JP5489872B2 - Claim structure information generation apparatus, claim structure information generation method, and program - Google Patents

Description

  The present invention relates to a claim structure information generating apparatus and the like for generating claim structure information which is information related to the structure of a claim.

  The claims described in “Claims”, which are the subject documents of patent applications, are usually written in a long and unique description style, which makes it extremely difficult for non-experts to read. Yes. On the other hand, a method of performing structural analysis using some fixed expressions often used in the description of actual patent claims has been proposed (for example, Patent Document 1, Patent Document 2, Non-Patent Document). 1). Further, a reduced expression method for easily grasping and storing the contents described in the claims has been proposed (see, for example, Patent Document 3).

Japanese Patent No. 3908261 Japanese Patent No. 4289871 JP 2004-240604 A

Akihiro Niimori, Manabu Okumura, Yuzo Marukawa, Makoto Iwayama, “Structural analysis of patent claims using clue phrases”, Transactions of Information Processing Society of Japan, Vol. 45, no. 3, p. 891-905, 2004

  However, Patent Documents 1 and 2 and Non-Patent Document 1 described above analyze the top-level structure of the claims, and no structural analysis was performed on the claims having a deeper structure. For example, when the proposed method is used for the following claims, the structural analysis is performed as shown in FIG. 16, and “the duct connected to the outside and the air blowing means for discharging the air” are performed. And a cleaning means for cleaning the air, and no analysis was performed on the portion "."

[Claims]
In the image forming apparatus,
A duct connected to the outside;
Air blowing means for discharging air;
A cleaning means for cleaning air,
The duct has an internal duct;
The cleaning means is installed in the internal duct,
An image forming apparatus.

  Further, the above-mentioned Patent Document 3 extracts “elements constituting the invention” and deletes the explanation thereof, and does not analyze the structure of the claims.

  The present invention has been made in accordance with the above situation, and an object of the present invention is to provide a claim structure information generation device and the like that can perform structural analysis on a claim having a deeper structure. To do.

  To achieve the above object, a claim structure information generating apparatus according to the present invention generates claim structure information that generates claim structure information indicating a description fragment in which a text of a claim is delimited and a structure of the description fragment. A storage unit for storing claim information that is a text of a claim, a morpheme analysis unit that performs morpheme analysis on the claim information, and a clause that specifies a position where a clause of the claim information is separated A surface layer delimiter information storage unit for storing two or more surface layer delimiter information having a specific part, surface layer clue information indicating delimitation of description fragments and indicating a relationship between the description fragments, and a token corresponding to the surface layer clue information; In the claim information, at a position corresponding to the clue information on the surface layer, a token granting unit that assigns a token corresponding to the clue information on the surface layer, and a clause that delimits the description fragment Corresponding to part-of-speech clue information in the phrase-type correspondence information storage unit for storing at least two clause-type correspondence information having clue information on lyrics and a segmental clause type corresponding to the part-of-speech clue information, and claim information The structure of a claim defined in advance by using a delimiter clause type assigning section for assigning a delimiter clause type corresponding to the clue information of the part of speech to the clause, and a token and a delimiter clause type given to the claim information In accordance with the rules for indicating the claim structure, a generation unit that generates the claim structure information indicating the structure of the description fragment of the claim claim information, and an output unit that outputs the claim structure information generated by the generation unit is there.

  With such a configuration, the structure of the description fragment included in the claim can be detected by using the clue information of the surface layer and the part of speech and the rule indicating the structure of the claim. Further, by performing analysis in units of phrases, it becomes possible to perform analysis at a more detailed level than in the case where analysis is performed for each morpheme as in the conventional example. As a result, a deep hierarchical structure can be analyzed.

  The claim structure information generating apparatus according to the present invention further includes a delimiter clause type changing unit that deletes a delimiter clause type existing at a position that is not a delimiter of the description fragment in accordance with the delimiter of the description fragment, and the generation unit includes: Claim structure information may be generated after the delimiter clause type changing unit deletes the delimiter clause type.

  With such a configuration, the delimiter clause type can be changed in accordance with the arrangement of descriptive fragments, and more appropriate structural analysis can be performed. For example, only the detection of the break position using the surface layer and part-of-speech clue information can also detect the formal break position, but this break phrase type change section should delete the break phrase type. Thus, an inappropriate break position can be deleted, and a more accurate analysis can be realized.

  In the claim structure information generating device according to the present invention, the delimiter clause type change unit includes a delimiter clause type that is a delimiter clause type at the end of the parallel descriptive fragment, and a description fragment that is in parallel before the terminal delimiter clause type. When there is a terminal delimiter clause type between a delimiter clause type that can be continued, a terminal delimiter clause type existing between both delimiter clause types may be deleted.

  With such a configuration, it is possible to delete a short break terminal delimiter clause type included in one description fragment and to make the delimiter position appropriate. For example, a claim is provided with a description “characterized by ………,…, ………,…”, “…,…,…,…, If there is a description of “”, even if “……” is included in the part of “……”, it is a short break in “……”, “...”, etc. Since it is considered not to be a delimiter of the description fragment, the terminal delimiter clause type corresponding to “Yes” is deleted.

In the claim structure information generating device according to the present invention, the delimiter clause type changing unit deletes the delimiter clause type when there is a newline and there is a delimiter clause type related to a verb not followed by a newline. May be.
With such a configuration, it is possible to delete a delimiter clause type related to a verb that is not followed by a new line that is a strong delimiter, and an appropriate delimiter position.

  In the claim structure information generation device according to the present invention, the generation unit has rules relating to tokens and delimiter clause types described in a context-free grammar, and an action for assigning a structure corresponding to the rules to the claim information. Claim structure information corresponding to the claim information may be generated using a parser generated using a grammar file.

  With such a configuration, by using a syntax analyzer generated by using a grammar file, the claim structure information obtained by analyzing the structure of the claim from the claim information to which the token or the demarcation clause type is assigned. Can be generated.

  Further, in the claim structure information generating apparatus according to the present invention, the description in the context free grammar of the structure of the claim is described in the direction from the end to the beginning of the claim, and the generation unit includes the token and delimiter clause type. Claim structure information may be generated by fetching the appended claim information in the direction from the end to the beginning.

  In the claims, there is a typical expression at the end, and the structure is determined in the direction from the end to the beginning. Therefore, such a configuration can avoid the occurrence of problems caused by grammatical ambiguity. sell. For example, ambiguities such as shift / reduce conflict and reduce / reduce conflict can be avoided.

In the claim structure information generating apparatus according to the present invention, the generating unit is described using a markup language by adding a mark indicating a structure corresponding to a rule included in the grammar file to the claim information. Claim structure information may be generated.
With this configuration, the claim structure information can be described using a markup language such as XML, for example.

Further, in the claim structure information generating apparatus according to the present invention, the generating unit includes a token granting unit according to a dependency relationship between a description fragment to which a predetermined token is assigned and a description fragment to which a delimiter clause type is assigned. Even if the dependency relationship between the description fragments indicated by the delimiter is specified by the token given by the delimiter and the delimiter clause type given by the delimiter clause type assignment unit, the claim structure information indicating the dependency relationship between the description fragments may be generated. Good.
With such a configuration, it is possible to obtain claim structure information indicating a dependency relationship between description fragments whose delimiter positions are indicated by tokens or delimiter clause types.

In the claim structure information generating apparatus according to the present invention, the output unit may display the claim structure information generated by the generation unit so that the relationship between the description fragments can be understood.
With such a configuration, by viewing the displayed claim structure information, it is possible to easily grasp the structure of a claim even if it is not a patent expert.

In the claim structure information generating device according to the present invention, the token includes a token corresponding to the inclusion relation of the component, and the delimiter clause type includes a parent of the token indicating the inclusion relation of the component. There are a parent delimiter clause type that is a delimiter clause type according to the component, and a child delimiter clause type that is a delimiter clause type according to the child component for the token indicating the inclusion relationship of the component, The output unit may display both description fragments so that the parent-child relationship between the description fragment corresponding to the parent-delimited clause type and the description fragment corresponding to the child-delimited clause type can be understood.
With such a configuration, the parent-child relationship of the constituent elements included in the claims can be shown more clearly.

Further, in the claim structure information generating device according to the present invention, a parent-child relationship acquisition unit for acquiring a parent-child relationship of the constituent elements using the claim information to which a parent-delimited clause type and a child-delimited clause type are attached, and a parent-child A hierarchy level acquisition unit that acquires the hierarchy level of the component using the parent-child relationship of the component acquired by the relationship acquisition unit, and the output unit includes a parent according to the hierarchy level acquired by the hierarchy level acquisition unit. A description fragment corresponding to the delimiter clause type and the child delimiter clause type may be displayed.
With such a configuration, it is possible to display according to the hierarchy level, and for example, it is possible to easily grasp in which hierarchy a certain component exists.

  According to the claim structure information generating apparatus and the like according to the present invention, it is possible to perform structural analysis even for a claim having a deeper structure.

The block diagram which shows the structure of the claim structure information generation apparatus by Embodiment 1 of this invention The flowchart which shows operation | movement of the claim structure information generation apparatus by the embodiment The figure which shows an example of the surface layer division | segmentation information in the embodiment The figure which shows an example of the phrase type corresponding | compatible information in the embodiment The figure which shows an example of the state transition table in the embodiment The figure which shows an example of the state transition table in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the grammar file in the embodiment The figure which shows an example of the claim structure information in the embodiment The figure which shows an example of the claim structure information in the embodiment The figure which shows an example of the display of the claim structure information in the embodiment The figure which shows an example of the display of the claim structure information in the embodiment The figure which shows an example of the display of the claim structure information in the embodiment The block diagram which shows the other structure of the claim structure information generation apparatus by the embodiment The figure which shows an example of the parent-child relationship of the component in the embodiment The figure which shows an example of the display of the claim structure information in the embodiment The schematic diagram which shows an example of the external appearance of the computer system in the said embodiment The figure which shows an example of a structure of the computer system in the said embodiment. The figure which shows an example of the analysis result of the claim in the conventional example

  Hereinafter, a claim structure information generation device according to the present invention will be described using embodiments. In the following embodiments, components and steps denoted by the same reference numerals are the same or equivalent, and repetitive description may be omitted.

(Embodiment 1)
A claim structure information generation device according to Embodiment 1 of the present invention will be described with reference to the drawings. The claim structure information generation device according to the present embodiment performs structural analysis on a claim having a deeper structure by using the clue information of the surface layer and the part of speech.

  FIG. 1 is a block diagram showing a configuration of a claim structure information generating apparatus 1 according to this embodiment. A claim structure information generating apparatus 1 according to the present embodiment generates claim structure information indicating a description fragment in which a text of a claim is delimited and a structure of the description fragment, and includes a storage unit 11 The morphological analysis unit 12, the phrase specifying unit 13, the surface layer delimiter information storage unit 14, the token granting unit 15, the phrase type correspondence information storage unit 16, the delimiter phrase type providing unit 17, and the delimiter phrase type changing unit 18 A generation unit 19 and an output unit 20. The description fragment is a unit for analyzing the dependency relationship in the claims, and is composed of a plurality of adjacent characters. The description fragment may be, for example, a clause including a component, or may be a description related to the component, and may be expressed in a manner unique to a claim (for example, an expression indicating an inclusion relation of a component or a jepson form). It may be an expression indicating a certain thing, an expression indicating a citation relationship, or the like.

  The storage unit 11 stores patent claim information which is a text of a patent claim. The claim information may be information of one claim or information of a plurality of claims, for example. In the latter case, the claims may or may not be dependent. Moreover, the claim information may be, for example, the information of [Claims] in a patent gazette or a patent publication gazette, or may not be so. The claim information stored in the storage unit 11 is an object of structural analysis.

  The process in which the claim information is stored in the storage unit 11 does not matter. For example, information may be stored in the storage unit 11 via a recording medium, or claim information transmitted via a communication line or the like may be stored in the storage unit 11. Alternatively, the claim information input via the input device may be stored in the storage unit 11. The storage in the storage unit 11 may be temporary storage in a RAM or the like, or may be long-term storage. The storage unit 11 can be realized by a predetermined recording medium (for example, a semiconductor memory, a magnetic disk, an optical disk, etc.). It goes without saying that information other than the claim information may be stored in the storage unit 11. For example, information on the result of processing by each component described later (specifically, the result of morphological analysis, the tokens assigned, the phrase type, etc.) may be stored in the storage unit 11.

  The morphological analysis unit 12 performs morphological analysis on the claim information stored in the storage unit 11. By the morphological analysis, the text of the claim information can be divided into morphemes, and the part of speech information for each morpheme can be obtained. The morphological analysis system is, for example, “MeCab” developed through the joint research unit project of the Graduate School of Informatics of Kyoto University-Nippon Telegraph and Telephone Corporation, Communication Science Basic Research Laboratories, and the Nara Institute of Science and Technology Graduate University. “ChaSen” (http://chasen.naist.jp) is known. For example, in the case of “MeCab”, by specifying an output format parameter, each morpheme is displayed in one line, and “headline”, “part of speech character string”, “utilization form”, “headline (basic form) ) ”And“ part of speech in the second layer ”can be output separated by tabs. Note that “MeCab” is designed to analyze line by line, so if there is a line break in the text of the claim information, the morphological analysis is performed for each line, and the part of speech character indicating the line break The results may be concatenated after inserting a column (eg, “NEWLINE”, etc.). The result of the morpheme analysis by the morpheme analysis unit 12 may be stored in the storage unit 11 or may be stored in another recording medium (not shown).

  The phrase specifying unit 13 specifies the segment break position of the claim information. There is no limitation on the method for specifying the break position of this clause. For example, by dividing the morphemes that are the result of morphological analysis into clause units, the segment break position may be specified, or by using clue phrases and other information, the clause information can be identified from the claim information itself. A delimiter position may be specified. In the present embodiment, the former case will be described. In the former case, the phrase specifying unit 13 includes, for example, the state transition table shown in FIG. 5 and “part-of-speech character string”, “utilization form”, and “part-of-speech part 2” as the result of morphological analysis. Using, morphemes may be grouped into phrases. For example, when the morphemes A and B are arranged in this order, the “current” row corresponding to the morpheme A and the “next” column corresponding to the morpheme B in the state transition table of FIG. If the numerical value is “1”, there is no clause segmentation position between morphemes A and B (that is, morphemes A and B are included in the same clause), and the numerical value is “0”. If so, there will be a clause break position between morphemes A and B (that is, morphemes A and B will be included in different clauses). By performing this process for all two adjacent morphemes, the phrase specifying unit 13 can specify the segment break position. Needless to say, the phrase break position may be specified by other methods. For example, for two consecutive morphemes, rules describing whether or not both should be included in the same phrase may be listed, and the phrases may be collected by sequentially applying the rules. The phrase delimiter position is specified by, for example, storing information indicating the delimiter position of the phrase (for example, the number of characters from the head or the number of bytes according to the phrase delimiter position) in the storage unit 11 or a recording medium (not shown). It may be to accumulate, may be to insert information indicating the delimiter position at the location corresponding to the delimiter position of the clause between morphemes, each morpheme that is the result of the morpheme analysis, It may be grouped into phrase units, or may be other methods that can know the segment break position. In addition, the timing at which the phrase specifying unit 13 specifies the segment break position is after the morpheme analysis by the morpheme analysis unit 12 when the specification is performed using the result of the morpheme analysis. In addition, when the identification is performed using a token corresponding to the clue information on the surface layer described later, the phrase specifying unit 13 specifies the segment break position after the token granting unit 15 described later gives the token. May be. Further, when the identification is performed from the claim information itself, the phrase identification unit 13 may identify the segment break position at a timing unrelated to morphological analysis or token assignment.

  The surface layer delimiter information storage unit 14 stores two or more surface layer delimiter information. Here, the surface layer delimiter information is information including surface layer clue information indicating demarcation of the description fragments and indicating the relationship between the description fragments, and a token corresponding to the surface layer clue information. The surface layer means the character string itself. The “heading” in the above description of MeCab is the surface layer. Surface clue information is a so-called clue phrase. The surface layer clue information is usually described using a regular expression. The token is information for identifying clue information on the surface layer.

  The surface layer clue information may include, for example, information according to a Jepson format description (for example, “in”), or according to a characteristic description (for example, “characterizes”, etc.). May be present, or may be in accordance with the description of the inclusion relationship (for example, “has”, “has,” etc.) of the constituent elements, and the description of the citation relationship (for example, , “According to“ claim ”, etc.)) may exist. Corresponding to the surface layer clue information, there may be a Jepson token, a feature description token, a component inclusion token, a citation token, or the like. This clue information on the surface layer serves as a delimiter for description fragments. That is, the position immediately before and / or immediately after the position corresponding to the clue information on the surface layer is usually the delimiter position of the description fragment. Further, the relationship between the description fragments is detected based on the clue information on the surface layer. That is, the clue information on the surface layer indicates the relationship between the description fragments in the sense that the relationship between the description fragments can be detected by using the clue information on the surface layer. The relationship between the description fragments is, for example, a dependency relationship of description fragments, and details of the dependency will be described later.

  The process in which two or more surface layer division | segmentation information is memorize | stored in the surface layer division | segmentation information storage part 14 does not ask | require. For example, two or more surface layer delimiter information may be stored in the surface layer delimiter information storage unit 14 via a recording medium, and two or more surface layer delimiter information transmitted via a communication line or the like may be stored in the surface layer delimiter information. It may be stored in the storage unit 14, or two or more surface layer delimiter information input via the input device may be stored in the surface layer delimiter information storage unit 14. The storage in the surface layer delimiter information storage unit 14 may be temporary storage in a RAM or the like, or may be long-term storage. The surface layer delimiter information storage unit 14 can be realized by a predetermined recording medium (for example, a semiconductor memory, a magnetic disk, an optical disk, or the like).

  In addition, each information included in the surface layer delimiter information, for example, surface layer clue information and tokens, may be the information itself, or may be information that can identify the information. In the latter case, the surface layer clue information and the like included in the surface layer separation information may be, for example, a pointer or an address indicating an area in which the surface layer clue information or the like is stored. This is the same for other information.

  In the claim information, the token granting unit 15 grants a token corresponding to the surface clue information to a position corresponding to the surface clue information. The method of granting a token is not limited as long as the position corresponding to the token can be identified as a result in the claim information. For example, the token granting unit 15 may add a token to the character string itself of the claim information, or store information associating the token with the range of the claim information corresponding to the token. It may be stored in the unit 11 or other recording medium (not shown). In the present embodiment, the former case will be described. For example, the token granting unit 15 may search the patent claim information using the surface clue information included in the surface layer delimiter information as a search key, and may give a token corresponding to the surface clue information to the hit location. . The token granting unit 15 may perform such processing for all the surface layer clue information included in the surface layer separation information. If the first surface clue information (for example, “characteristically”) includes the second surface clue information (for example, “characteristically”), the token granting unit 15 For example, the token may be assigned by searching first from the clue information on the first surface layer, and the portion to which the token is assigned may not be searched during subsequent searches. Alternatively, for example, the token granting unit 15 gives a token by performing a search first from the clue information on the second surface layer, and when the search is performed using the clue information on the first surface layer, the token is already assigned. A search may also be performed for a spot where a token has already been assigned, and a token may be assigned by overwriting when a hit has already been made.

  The phrase type correspondence information storage unit 16 stores two or more phrase type correspondence information. Here, the phrase type correspondence information is information having clue information on the part of speech of the clause that is a delimiter of the description fragment and a break phrase type corresponding to the clue information on the part of speech. The part-of-speech clue information can be said to be a part-of-speech version of a surface layer clue phrase. That is, the part-of-speech clue information is clue information including at least a condition related to the part-of-speech. The part-of-speech clue information may be information that identifies only the part of speech, or information that identifies the part of speech and the inflected form. In addition to the part of speech or the part of speech and inflected form, the surface information is also identified. It may be information, or information related to other parts of speech. The part-of-speech clue information may include part-of-speech part-of-speech information that is not related to the segmentation of the description fragment. In that case, the phrase type corresponding to the part-of-speech clue information of the phrase unrelated to such a break and the break phrase type are simply referred to as a phrase type. That is, the clause type includes a delimiter clause type in which the clause serves as a delimiter for description fragments, and a clause type unrelated to the delimiter.

  The part-of-speech clue information includes, for example, information corresponding to the delimiter clause type at the end of parallel description fragments (hereinafter also referred to as “terminal delimiter clause type”) (for example, “end of verb or auxiliary verb”) Depending on the delimiter clause type (hereinafter sometimes referred to as the “continue delimiter clause type”) in which parallel description fragments can continue (that is, cannot be said to be terminal). May be present (for example, “verb or auxiliary verb continuous form”, “noun + and”, etc.). Further, other part-of-speech cue information may exist.

  The process in which two or more phrase type correspondence information is stored in the phrase type correspondence information storage unit 16 does not matter. For example, two or more phrase type correspondence information may be stored in the phrase type correspondence information storage unit 16 via a recording medium, and two or more phrase type correspondence information transmitted via a communication line or the like may be stored. The phrase type correspondence information storage unit 16 may store the phrase type correspondence information, or two or more phrase type correspondence information input via the input device may be stored in the phrase type correspondence information storage unit 16. May be. Storage in the phrase type correspondence information storage unit 16 may be temporary storage in a RAM or the like, or may be long-term storage. The phrase type correspondence information storage unit 16 can be realized by a predetermined recording medium (for example, a semiconductor memory, a magnetic disk, an optical disk, or the like).

  The demarcation phrase type assigning unit 17 assigns a demarcation phrase type corresponding to the part-of-speech clue information to the phrase corresponding to the part-of-speech clue information in the claim information. The delimited phrase type assigning unit 17 may assign a phrase type that is not a delimited phrase type. In addition, the demarcation phrase type assigning unit 17 performs the assigning process using the result of the morpheme analysis by the morpheme analyzing unit 12. The granting of the phrase type is not limited as long as the position corresponding to the phrase type can be identified as a result in the claim information, as in the token grant described above. For example, the clause type may be added to the character string itself of the claim information, or information that associates the clause type with the scope of the claim information corresponding to the clause type is stored in the storage unit 11 or You may accumulate | store in another recording medium which is not illustrated. In the present embodiment, the former case will be described. For example, the delimited phrase type assigning unit 17 searches the result of the morphological analysis using the clue information of the part of speech included in the phrase type correspondence information as a search key, and assigns the phrase type corresponding to the clue information of the part of speech to the hit location. May be. The delimited phrase type assigning unit 17 may perform such processing for clue information of all parts of speech included in the phrase type correspondence information. When the first part-of-speech clue information includes the second part-of-speech clue information, for example, the phrase type is given first from the first part-of-speech clue information, as in the case of token grant. The phrase type may be added to the part where the phrase type has already been assigned, or the phrase type may be assigned first from the clue information of the second part of speech. When the phrase type is assigned using the part-of-speech clue information, the phrase type may be assigned by overwriting the part where the phrase type has already been assigned. As a result of the provision of the phrase type by the delimiter phrase type assignment unit 17, one phrase type is assigned to one phrase. There may be clauses to which no clause type is assigned. In the following description, the phrase to which the phrase type A is given may be simply referred to as A.

  The delimiter clause type changing unit 18 deletes delimiter clause types that exist at positions that are not delimiters of description fragments, in accordance with the delimitation of description fragments. “According to the arrangement of descriptive fragments” refers to the arrangement of tokens and delimiter clause types assigned to the claim information. That is, when the sequence of the given tokens or delimiter clause types satisfies a predetermined condition, the delimiter clause type change unit 18 deletes a specific delimiter clause type in the sequence that satisfies the condition. Specifically, the delimiter clause type changing unit 18 terminates between the end delimiter clause type and a continuation delimiter clause type that is earlier than the end delimiter clause type (that is, located at a position closer to the head). If there is a delimiter clause type, the terminal delimiter clause type existing between the terminal delimiter clause type and the continuation delimiter clause type may be deleted. Further, the delimiter clause type changing unit 18 is a delimiter clause type related to a verb in which a newline is present and no newline follows (this delimiter clause type may be a terminal delimiter clause type related to a verb, for example, If there is a continuation break clause type for a verb), the break clause type may be deleted. The delimiter clause type that “no line feed follows” is a delimiter clause type that does not have a line break immediately after the clause to which the delimiter clause type is assigned. The “delimiter clause type related to a verb” is a delimiter clause type corresponding to clue information of a part of speech related to a verb or auxiliary verb. Moreover, the demarcation phrase type changing unit 18 may perform the processing within a range of interest. For example, the delimiter clause type changing unit 18 performs this process on a portion closer to the head than the position where the Jepson token is given (this portion may be referred to as the “first half”). This processing may be performed for a portion closer to the end than the position where the format token is assigned (this portion may be referred to as “second half portion”). Further, for example, the demarcation clause type changing unit 18 may perform this process for the entire claim. In addition, the delimiter clause type changing unit 18 may change (change) the phrase type in addition to deleting the delimiter clause type. The delimiter clause type changing unit 18 may change a clause type that is not a delimiter clause type to a delimiter clause type, or may change a delimiter clause type to a clause type that is not a delimiter clause type. The latter can be said to be a deletion of the delimiter clause type in the sense that the delimiter clause type is eliminated.

  Further, the delimiter clause type changing unit 18 may change a clause type that is not a delimiter clause type to a delimiter clause type in accordance with the arrangement of tokens and clause types. In this case, a new delimiter clause type is generated. For example, the delimiter clause type change unit 18 has a new line, and when there is a phrase type related to a verb followed by a new line (this phrase type is not a delimited phrase type), the phrase The type may be changed to a delimited clause type. Also, for example, when there is a phrase type related to a noun immediately before the token of the inclusion relation of the component (assuming that this phrase type is not a separator phrase type), the delimiter phrase type changing unit 18 changes the phrase type. You may change to a delimited clause type. The phrase type related to the noun is a phrase type corresponding to the clue information of the part of speech related to the noun. In addition, the phrase type change by the delimiter phrase type changing unit 18 includes a change to make the description match the analysis when the generation unit 19 described later performs an analysis using a syntax analyzer. May be.

  Here, a brief description will be given of token assignment, clause type assignment, and clause type change. By adding a token or clause type to the claim information, it becomes possible to specify the delimiter position of the description fragment in the claim information. The token is given using a cue phrase on the surface layer, and the phrase type is given using a condition related to the part of speech. In addition, when the phrase type is changed, the delimiter position becomes more appropriate. Since the claim structure information generating apparatus 1 according to the present embodiment detects the delimiter position of the description fragment in units of clauses, the delimiter position of the descriptive fragment is detected in units of morphemes as in the conventional example. Even more detailed rules regarding the structure can be set. In the case of the conventional example, since the unit is a morpheme finer than a clause, when trying to set a rule related to a more detailed structure, the rule becomes very fine, and the rule is substantially created. I couldn't. On the other hand, since the claim structure information generating apparatus 1 according to the present embodiment uses clauses as a unit, it is possible to set rules related to a more detailed structure within a reasonable range, and to analyze a more detailed structure. It was.

  The generation unit 19 indicates the structure of the description fragment of the claim information in accordance with the rule indicating the structure of the patent claim, which is specified in advance, using the token and the delimiter clause type given to the claim information. Claim structure information is generated. When the delimiter clause type change unit 18 deletes the delimiter clause type or changes the phrase type, the generation unit 19 deletes the delimiter clause type or the phrase type is changed by the delimiter clause type change unit 18. After the change, the claim structure information is generated. That is, the generation unit 19 generates the claim structure information using the token and the delimiter clause type given to the claim information after the deletion or change. The claim structure information is information indicating at least a description fragment delimiter and an attribute of the description fragment. The attribute of the description fragment may indicate, for example, that the description fragment includes the component name, or may indicate that the description fragment is an explanation for the description fragment including the component name. Well, it may indicate that the description fragment is a description in the Jepson format, may indicate that the description fragment is a description of the inclusion relation of the constituent elements, and the dependency relationship between the description fragments. May be shown. The generation unit 19 can specify the delimitation position of the description fragment according to the type of token or delimiter clause type given to the claim information. For example, the generation unit 19 can specify the delimitation position of the description fragment using information including a token or delimiter clause type and a delimiter position corresponding to the token or delimiter clause type. In addition, the generation unit 19 can specify the attribute of the description fragment according to the token or the demarcation clause type itself given to the claim information or the arrangement thereof. For example, the generation unit 19 can specify the attribute of the description fragment by using information including a token or a delimiter clause type and an attribute corresponding to the token or delimiter clause type. For example, the generation unit 19 can specify the attribute of the description fragment by using information including information indicating a sequence of tokens and delimiter clause types and attributes corresponding to the token and delimiter clause types.

  The generation unit 19 may generate the claim structure information corresponding to the claim information to which the token and the demarcation clause type are given, for example, using a syntax analyzer. The parser (parser) was generated using a grammar file having rules for token and delimiter clause types described in a context-free grammar, and an action for assigning a structure corresponding to the rules to claim information. Is. That is, the claim structure information is obtained by analyzing the claim information to which the token and the demarcation clause type are given by the parser. In this case, the generation unit 19 generates the claim structure information in accordance with a grammar file that is a rule indicating the structure of a patent claim defined in advance. The generating unit 19 attaches a mark (for example, a tag) indicating a structure corresponding to a rule included in the grammar file to the claim information, and claims structure information described using a markup language May be generated. The rule may be, for example, a rule related to a sequence of tokens and delimited clause types. The action may be in accordance with the attribute of the description fragment. The markup language may be, for example, XML (Extensible Markup Language) or any other. As the XML structure, PML (Patent Markup Language) proposed by Hidekazu Tanigawa et al. May be used (http://www.ird-pat.com/pml.html). Further, when generating the claim structure information using the parser, the generation unit 19 reads the claim information to which the token and the delimiter clause type are added in the direction from the end to the start. Structural information may be generated. In the claims, there is a typical expression at the end, and the structure is determined in the direction from the end to the beginning. Therefore, such a configuration can avoid the occurrence of problems caused by grammatical ambiguity. sell. For example, ambiguities such as shift / reduce conflict and reduce / reduce conflict can be avoided. In this case, it is assumed that the description in the context-free grammar of the structure of the claim is described in the direction from the end to the beginning of the claim.

  In addition, the generation unit 19, for example, determines the token and delimiter clause given by the token grant unit 15 according to the dependency relationship between the description fragment to which the predetermined token is given and the descriptive clause to which the descriptive clause type is given. The dependency structure between the description fragments indicated by the delimiter clause type provided by the type assigning unit 17 may be specified, and the claim structure information indicating the dependency relationship between the description fragments may be generated. In this case, the dependency relationship between the description fragment to which the token is specified in advance and the description fragment to which the delimiter clause type is assigned is a rule indicating the structure of the claimed claim. The dependency relationship between the description fragment to which the token is assigned and the description fragment to which the delimiter clause type is assigned may include, for example, the dependency relationship between the description fragments to which the token is attached, It may include a dependency relationship between descriptive fragments with a delimited clause type, or include a dependency relationship between a descriptive fragment with a token and a descriptive clause type. Good. For example, when it is preliminarily specified that the description fragment to which the delimiter clause type A is assigned relates to the description fragment to which the token B is given, the generation unit 19 describes the description fragment to which the delimiter clause type A is assigned. May generate claim structure information indicating a dependency relationship relating to the description fragment to which the token B is assigned. The claim structure information may be, for example, information in which a description fragment is associated with information for identifying the description fragment related to the description fragment.

  The output unit 20 outputs the claim structure information generated by the generation unit 19. Here, the output may be, for example, display on a display device (for example, a CRT or a liquid crystal display), transmission via a communication line to a predetermined device, printing by a printer, or output to a recording medium. It may be accumulated or delivered to another component. When the output is a display, the output unit 20 may display the claim structure information generated by the generation unit 19 so that the relationship between the description fragments can be understood. For example, when the claim structure information indicates the dependency relationship between the description fragments, the output unit 20 may display the description fragment and the graphic corresponding to the dependency between the description fragments. Good. The figure corresponding to the dependency may be, for example, a line figure connecting two description fragments having a dependency relationship, and the same mark given to both of the two description fragments having a dependency relationship. The figure may also be a figure showing other dependency relationships. Further, for example, when the claim structure information is information indicating at least a description fragment delimiter and an attribute of the description fragment, the output unit 20 responds to a predetermined dependency relationship of attributes. , It may be displayed so that the dependency relationship between the description fragments can be understood. The display may display, for example, a description fragment and a graphic corresponding to the dependency between the description fragments as described above.

  The output unit 20 may or may not include an output device (for example, a display device or a printer). The output unit 20 may be realized by hardware, or may be realized by software such as a driver that drives these devices.

  Note that any two or more storage units of the storage unit 11, the surface layer delimiter information storage unit 14, and the phrase type correspondence information storage unit 16 may be realized by the same recording medium, or may be separate recording media. It may be realized with. In the former case, for example, the area storing the claim information becomes the storage unit 11, and the area storing the surface layer delimiter information becomes the surface layer delimiter information storage unit 14.

  Next, operation | movement of the claim structure information generation apparatus 1 by this Embodiment is demonstrated using the flowchart of FIG. Note that, in this flowchart, a description will be given of a case where all the claims information subjected to morphological analysis, token assignment, and the like are stored in the storage unit 11.

  (Step S <b> 101) The morpheme analysis unit 12 performs morpheme analysis on the claim information stored in the storage unit 11. Then, the morphological analysis unit 12 accumulates the result in the storage unit 11. The result of the morphological analysis may be, for example, a text file in which information on one morpheme is on a single line.

  (Step S <b> 102) The token grant unit 15 grants a token to the claim information stored in the storage unit 11 using the surface layer delimiter information stored in the surface layer delimiter information storage unit 14. The given token is stored in the storage unit 11 in association with the corresponding part of the claim information. For example, the token may be added to the same line as the portion corresponding to the clue information on the surface layer in the text file as a result of the morphological analysis. In addition, when the location corresponding to the clue information on the surface layer extends over a plurality of morphemes, the token granting unit 15 may combine the plurality of morphemes into one, for example.

  (Step S <b> 103) The phrase specifying unit 13 specifies phrase breaks by collecting the morphemes after morpheme analysis stored in the storage unit 11. The phrase specifying unit 13 may, for example, add information indicating that the phrase is at a phrase break position between morphemes that are the phrase break positions specified.

  (Step S104) The demarcation phrase type giving unit 17 uses the phrase type correspondence information stored in the phrase type correspondence information storage unit 16 for the clause of the claim information stored in the storage unit 11, and Give clause type. As described above, the phrase type to be added includes the delimiter phrase type. The assigned phrase type is stored in the storage unit 11 in association with the corresponding phrase of the claim information. For example, the phrase type may be added to the same line as the part corresponding to the part-of-speech clue information in the text file after specifying the segment break position. When one clause is composed of a plurality of morphemes, for example, the delimiter clause type assigning unit 17 summarizes the plurality of morphemes so that one clause is provided for each clause after the clause type is assigned. May be.

  (Step S105) The delimited phrase type changing unit 18 changes the assigned phrase type according to the sequence of tokens and phrase types. As described above, the change includes deleting a delimiter clause type that exists at a position that is not a delimiter of the description fragment. Further, as described above, the delimiter clause type changing unit 18 may change a clause type that is not a delimiter clause type to a delimiter clause type, or may change other clause types.

  (Step S106) The generation unit 19 generates claim structure information from the claim information to which the token or clause type is assigned.

  (Step S107) The output unit 20 outputs the claim structure information generated by the generation unit 19. Then, a series of processes for generating and outputting the claim structure information is completed.

  In addition, when a plurality of claims are included in the claim information, the processing of steps S101 to S107 may be repeatedly executed for each claim in the flowchart of FIG. 2, or a plurality of claims. You may perform the process of step S101-S107 with respect to the whole. In addition, when the token given in specifying the segment break position is not used, the token granting process may be performed at any time before the phrase type change process (step S105).

  Next, operation | movement of the claim structure information generation apparatus 1 by this Embodiment is demonstrated using a specific example. In this specific example, a process related to one claim will be described. However, when a plurality of claims are included in the claim information, a plurality of claims can be obtained by repeatedly executing the process. Analysis can be performed. Also here, as in the description of the conventional example, it is assumed that the following claim information is stored in the storage unit 11.

[Claim Information]
In the image forming apparatus,
A duct connected to the outside;
Air blowing means for discharging air;
A cleaning means for cleaning air,
The duct has an internal duct;
The cleaning means is installed in the internal duct,
An image forming apparatus.

Further, in this specific example, it is assumed that the surface layer delimiter information storage unit 14 stores six types of surface layer delimiter information shown in FIG. In addition, in the expression of the clue information on the surface layer in FIG. 3 and the subsequent description, the regular expression notation of Perl language is used. For the Perl language, refer to the following document.
Literature: Larry Wall, Tom Christiansen, Randal L. Shwartz, "Programming Perl Revised Edition", O'Reilly Japan, 1997

  In the surface layer clue information of the surface layer delimiter information in FIG. 3, the clue information described in a plurality of lines means that a corresponding token is given when any one of them corresponds. The token “JEPSON_CUE” corresponds to the clue phrase characteristic of the Jepson-type claim. The tokens “FEATURE_CUE” and “FEATURE_K_CUE” correspond to the cue phrase described in the claims. The token “COMPOSE_K_CUE” corresponds to the clue phrase used in the claim of the component enumeration type “XX with ˜, ˜, and ˜”, and the token “COMPOSE_R_CUE” , And, and corresponds to the clue phrase used in the claim of the modified component enumeration type of “. The token “QUIOTE_CUE” corresponds to the clue phrase described in the claims.

  In this specific example, it is assumed that the phrase type correspondence information storage unit 16 stores the phrase type correspondence information shown in FIG. In the part-of-speech clue information in FIG. 4, “X” is a noun or a symbol. Symbols include punctuation marks. “Y” is a verb or prefix of “body connection special 2”. Also, in the part-of-speech clue information of the phrase type correspondence information in FIG. 4, when the clue information described in a plurality of lines corresponds to any one, it means that the corresponding phrase type is given. . For example, a phrase type “VPR” is given to a clause that ends with a verb or auxiliary verb, or a basic form of a verb or auxiliary verb followed by a phrase that ends with “with” or “with” . In addition, the phrase types in FIG. 4 include both delimited phrase types and non-delimited phrase types, but VPR_P, VPK_P, and NPT_P are delimited phrase types indicating delimiters of description fragments. VPK_P is a terminal delimiter clause type, and VPR_P and NPT_P are continuation delimiter clause types.

  First, the morpheme analysis unit 12 performs morpheme analysis for each line of the claim information stored in the storage unit 11 (that is, the morpheme analysis is performed seven times), and at the position of the line break “NEWLINE” was inserted (step S101). Here, the morphological analysis system “MeCab” was used. The results are as follows. Note that the result of the morphological analysis by the morphological analysis unit 12 is stored in the storage unit 11. In the following results, as described above, for each morpheme, five pieces of information of “heading (surface layer)”, “part of speech character string”, “utilization form”, “heading (basic form)”, and “part of speech in the second layer” Is obtained.

[Results of morphological analysis]
Image noun image general formation noun formation sa change connection device noun device in sa change connection particle in case particle, symbol, punctuation NEWLINE
External noun external in general case particle connected to case particle connected noun connected saber connected verb connected to self-relative self-standing verb basic formed suffix duct noun duct general and particle and parallel particle, symbol, punctuation NEWLINE
Air noun air in general particle in case particle discharge noun discharge insanely connected verb basic form self-sustained noun blast insane connecting means noun means general and particle and parallel particle, symbol, punctuation NEWLINE
Air noun air general particle as case particle cleansing noun purifier saber conjunctive nounization suffix verb basic form self-standing clean noun purifier savariant conjunctive nounization suffix means noun means general and particles and parallel particles, symbols and punctuation as particles Prepared Verb Consistent Provided Independence, Symbols, Readings NEWLINE
The above-mentioned nouns The above-mentioned change connection ducts Nouns Ducts General particles are co-inners Internal nouns Internal general ducts Nouns Ducts General particles have case particles Verbs Conjunctive forms Independence, Symbols, Readings NEWLINE
The above noun The above change connection cleanup Noun Clean The change connection Nounization Suffixing means Noun Meaning Generally, the particle is the particle, symbol, punctuation The above noun The inside change of the internal connection Noun Internal duct Noun Duct General internal noun In the suffix Noun Installation Sabari Connected Verb Naturally Connect Freestanding Verb Conjunctive Form Suffix Particles Conjunctive Verb Verbs Basic Forms Non Independent, Symbols, Readings NEWLINE
That noun that is independence particle is a case particle feature Noun feature General and particle and case particle Verb Basic form Autonomous image Noun image General formation Noun formation Change connection device Noun device Change connection. Sign. Punctuation

  Next, the token granting unit 15 grants a token to the result of the morphological analysis using the surface layer delimiter information shown in FIG. 3 stored in the surface layer delimiter information storage unit 14 (step S102). Specifically, the token assigning unit 15 performs a search in the heading (surface layer) in the result of morphological analysis using the surface clue information as a search key, and assigns a token corresponding to the surface clue information to the hit. . If the token is to be assigned to two or more morphemes, the tokens are given to the two or more morphemes as one set (chunk). Further, the token granted by the token granting unit 15 is also stored in the storage unit 11. In this case, tokens “JEPSON_CUE”, “COMPOSE_R_CUE”, and “FEATURE_K_CUE” are given.

[Result of token grant]
Image noun image general formation noun formation change connection device noun device change connection, JEPSON_CUE
NEWLINE
External noun external in general case particle connected to case particle connected noun connected saber connected verb connected to self-relative self-standing verb basic formed suffix duct noun duct general and particle and parallel particle, symbol, punctuation NEWLINE
Air noun air in general particle in case particle discharge noun discharge insanely connected verb basic form self-sustained noun blast insane connecting means noun means general and particle and parallel particle, symbol, punctuation NEWLINE
Air noun Air in general particle as case particle cleanup noun purifier saber conjunctive nounization suffix verb basic form self-standing clean noun purifier savariant conjunctive nounization suffix means noun means in general, particle, parallel particle, symbol, punctuation, COMPOSE_R_CUE
NEWLINE
Said noun said sagittal connection duct noun duct general particle is a particle inner noun internal general duct noun duct general, COMPOSE_R_CUE
NEWLINE
The above noun The above change connection cleanup Noun Clean The change connection Nounization Suffixing means Noun Meaning Generally, the particle is the particle, symbol, punctuation The above noun The inside change of the internal connection Noun Internal duct Noun Duct General internal noun In the suffix Noun installation Sabari connected Verb naturally connected to self-supporting verb Consecutive form Suffix particle and connected particle are verb Basic form is non-independent, symbolic and punctuation mark FEATURE_K_CUE
Image noun image general formation noun formation change connection device noun device change connection. Sign. Punctuation

  Thereafter, the phrase specifying unit 13 is shown in FIG. 5A and FIG. 5B that are stored in a recording medium (not shown) and information of “part of speech character string”, “utilized form”, and “second part of speech” of each morpheme. Using the state transition table, the segment break position is specified (step S103). At that time, the phrase specifying unit 13 does not specify the position where the token is delimited for the part to which the token is given. Therefore, the part where the token is given and the part adjacent to it may be different from the division of the grammatical clause. In this specific example, it is assumed that the phrase specifying unit 13 gives information indicating that it is before the break to the morpheme immediately before the break position of the phrase. The specific result of the phrase delimiter position by the phrase specifying unit 13 is as follows. As a result, the information “」 ”indicating that the morpheme is immediately before the break position is assigned to the morpheme immediately before the break position of the phrase. In addition, information indicating the break position of the phrase specified by the phrase specifying unit 13 is also stored in the storage unit 11.

[Result of specifying the position of clause break]
Image noun image general formation noun formation change connection device noun device change connection ◎
In JEPSON_CUE ◎
NEWLINE
External noun External In general particle to case particle ◎
Connected Noun Connected Versatile Connected Verb Connected to Self-standing Verb Basic Formed Suffix ◎
Duct Noun Duct General and particle and parallel particle, symbol, reading ◎
NEWLINE
Air noun air general particle as case particle ◎
Ejected noun Ejected Verb Connected verb Basic form Independent ◎
Blowing noun Blowing change connection means Noun means General and particle and parallel particle, symbol, reading ◎
NEWLINE
Air noun air general particle as case particle ◎
Clean Noun Cleanliness Change Connection Nounization Suffix ◎
To verb basic form to independence ◎
Cleansing noun cleansing change connection nounization suffix means noun means general and particle and parallel particle, symbol, reading ◎
COMPOSE_R_CUE ◎
NEWLINE
The above-mentioned noun The above-mentioned variable connection duct Noun duct In general, the particle is the particle ◎
Internal Noun Internal General Duct Noun Duct General ◎
COMPOSE_R_CUE ◎
NEWLINE
The above nouns The above change connection cleanup Nouns Clean up change connection Nounization Suffixing means Noun means In general, the particle is a particle, symbol, punctuation mark ◎
The above nouns The above-mentioned internal connection inside nouns Internal general ducts Nouns Ducts General internal nouns Internal suffixes to particles and case particles ◎
Installation noun installation change verb connected to self-relative self-supporting verb conjunctive form suffixed particle connected verbal verb verb basic form non-independent, symbol, punctuation mark ◎
FEATURE_K_CUE ◎
Image noun image general formation noun formation change connection device noun device change connection. Sign. Punctuation ◎

  Next, using the phrase type correspondence information shown in FIG. 4 stored in the phrase type correspondence information storage unit 16, the break phrase type assigning unit 17 selects a phrase type for the result of specifying the phrase break position. (Step S104). At that time, the delimited phrase type assigning unit 17 assigns the phrase type using one or more morphemes from the morpheme at the end of the phrase. Specifically, the delimited phrase type assigning unit 17 searches for a phrase corresponding to the part-of-speech clue information, and assigns the phrase type corresponding to the part-of-speech clue information to the hit phrase. Since the result of the morphological analysis is not used after the phrase type is given, information other than the heading (surface layer) may be deleted from the result of the morphological analysis. The result of the phrase type assignment is also stored in the storage unit 11. In this case, the phrase types “NP”, “PP”, “VPK”, and the like are given.

[Result of granting clause type]
Image forming device NP
In JEPSON_CUE
NEWLINE
PP outside
Connected VPK
Duct and NPT_P
NEWLINE
Air to NPO
VPK to discharge
Air blowing means, NPT_P
NEWLINE
Air to NPO
NP
VPK
Cleaning means and NPT_P
COMPOSE_R_CUE
NEWLINE
The duct is PP
Internal duct NP
COMPOSE_R_CUE
NEWLINE
The cleaning means is PP
PP inside the internal duct
VPK_P installed
FEATURE_K_CUE characterized by
Image forming apparatus. NP

  Thereafter, the delimiter clause type changing unit 18 changes the clause type (step S105). The specific method of the change is as follows. In the description of the method, the “noun group” is a series of nouns. The “noun group” may be a phrase to which the phrase type “NP” is assigned, or may be a series of nouns included in the assigned phrases such as the phrase types “NPT” and “NPT_P”. Further, in the following description, a clause to which a token or clause type is assigned is represented by the token or clause type itself for convenience of explanation. For example, “VPR_P” strictly means “a clause to which VPR_P is assigned”.

[Explanation of phrase type change]
1. When JEPSON_CUE exists, the following processing is performed.
(1) If any of FEATURE_CUE, FEATURE_K_CUE, COMPOSE_K_CUE exists in the first half (that is, the part from the beginning of the claims to immediately before JEPSON_CUE), the following processing A, processing B, and processing C are performed.
(Processing A) All clause types after FEATURE_CUE, FEATURE_K_CUE, or COMPOSE_K_CUE that appear last are replaced with HP.
(Processing B) VPR_P, NPT_P, COMPOSE_R_CUE, or processing range (for example, when processing of the first half is being performed in the part before the “noun group” immediately before COMPOSE_K_CUE, FEATURE_CUE, or the part before FEATURE_K_CUE VPK immediately before the “noun group” until the first half of the first half or the second half is processed, or the first half of the second half or the entire second when the entire process is performed. Alternatively, when VPK_P exists, or when VPK or VPK_P exists immediately before FEATURE_K_CUE, VP other than VPK or VPK_P immediately before “nouns” and other than VPK or VPK_P immediately before FEATURE_K_CUE, And K *, replace all the NP * PP.
(Process C) If VPR_P is encountered in Process B, this process is repeated. In the process B, when NPT_P is encountered, this process is repeated in a part before the “noun group” including the NPT_P. When the COMPOSE_R_CUE is encountered in the process B, this process is repeated in the part before the “noun group” immediately before. In the process B, when the head of the processing range is encountered, this process is terminated.
(2) When none of FEATURE_CUE, FEATURE_K_CUE, and COMPOSE_K_CUE exists in the first half, the following processing D, processing E, and processing F are performed.
(Process D) Recognize “noun group” immediately before JEPSON_CUE, and change all phrase types constituting it to HP.
(Processing E) If VPK or VPK_P exists immediately before HP before the HP encounters one of VPR_P, NPT_P, COMPOSE_R_CUE, or the head of the processing range before HP, HP VPK * and NP * are all replaced with PP except for VPK or VPK_P immediately before.
(Process F) When VPR_P is encountered in Process E, this process is repeated. In the process E, when NPT_P is encountered, this process is repeated in a part before the “noun group” including the NPT_P. In the process E, when COMPOSE_R_CUE is encountered, this process is repeated in the part before the “noun group” immediately before. In the process E, when the head of the processing range is encountered, this process is terminated.
(3) If NEWLINE exists in the first half, the next process G is performed.
(Process G) VPK followed by NEWLINE is changed to VPK_P, VPR followed by NEWLINE is changed to VPR_P, VPR_P not followed by NEWLINE is changed to VPR, and VPK_P not followed by NEWLINE is changed to VPK.
(4) If any of FEATURE_CUE, FEATURE_K_CUE, COMPOSE_K_CUE exists in the latter half (that is, the part immediately after JEPSON_CUE to the end of the claims), the following process H, process I, and process J are performed.
(Processing H) All phrase types after the last occurrence are replaced with BP. If none exists, the last “noun group” is recognized and all the phrase types constituting it are replaced with BP.
(Process I) The process is the same as the process B.
(Process J) This is the same process as Process C above.
(5) When none of FEATURE_CUE, FEATURE_K_CUE, and COMPOSE_K_CUE exists in the latter half, the following processing K, processing L, and processing M are performed.
(Processing K) Recognize the last “noun group” and change all the phrase types constituting it to BP.
(Process L) In the process E, the process of replacing HP with BP is performed.
(Process M) In the process F, the process of replacing HP with BP is performed.
(6) If NEWLINE exists in the latter half, the process G is performed.
2. When JEPSON_CUE does not exist, the processing performed in the latter half in the above (4) to (6) is performed once for the entire claim.
3. NPT immediately before COMPOSE_K_CUE and COMPOSE_R_CUE is changed to NPT_P, and NP immediately before COMPOSE_K_CUE and COMPOSE_R_CUE is changed to NPT_P.
4). Remove NEWLINE.

In the description of the phrase type change described above, in the process B and the process E, the process of replacing VPK_P (terminal delimiter phrase type) with PP is the terminal delimiter phrase type (VPK_P) and the continuation delimiter phrase type (VPR_P, This corresponds to the process of deleting the terminal delimiter clause type (VPK_P) existing between the NPT_P) and NPT_P). In process G, VPR_P that is not followed by NEWLINE is changed to VPR, and VPK_P that is not followed by NEWLINE is changed to VPK when there is a line break and the delimiter clause type (VPR_P) for a verb that does not follow a new line. , VPK_P) corresponds to the process of deleting the delimiter clause type.
The phrase type after the delimiter phrase type changing unit 18 changes as described above is as follows.

[Result of phrase type change]
Image forming device HP
In JEPSON_CUE
PP outside
Connected VPK
Duct and NPT_P
PP for air
VPK to discharge
Air blowing means, NPT_P
PP for air
Cleaning PP
VPK
Cleaning means and NPT_P
COMPOSE_R_CUE
The duct is PP
Internal duct NPT_P
COMPOSE_R_CUE
The cleaning means is PP
PP inside the internal duct
VPK_P installed
FEATURE_K_CUE characterized by
Image forming apparatus. BP

  Thereafter, the generation unit 19 generates claim structure information from the claim information to which the token and the phrase type are assigned (step S106). In the generation, in this specific example, the claim structure information is generated by using the parser (syntax analyzer) generated by the parser generator using the grammar file shown in FIG. . The grammar file in FIG. 6 is Parse :: Yapp (Francois Desarmenien, Parse-Yapp-1.05, http://search.cpan.org/dist/) which is a parser tool for Bison compatible Perl. This is a description example for Parse-Yapp /). Considering that there is usually a canonical expression at the end and the structure is fixed in the direction from the end to the beginning, the description in the context-free grammar is written in the description of the claims from the end to the beginning. Described by direction. In FIG. 6, the generated XML is stored as a character string in a variable called $ retval. When the result of the phrase type change is taken into the parser generated from the grammar file shown in FIG. 6, the claim structure information shown in the XML file in FIG. 7 is generated. At that time, the phrase type change result is converted into the direction from the end to the beginning and input to the parser generated from the context-free grammar. As the XML structure, the aforementioned PML was used.

  The output unit 20 reads the claim structure information of FIG. 7 and displays it on the display. As a result, the display showing the relationship between the description fragments according to the claim structure information shown in FIG. 8 is performed (step S107). By viewing this display, the user can easily grasp the structure of the claims. Note that a parser (parser) may also be used when generating the display shown in FIG. 8 from the XML file shown in FIG. This parser is a known parser that displays a structure described in a markup language.

  Here, a specific example of another process for generating the claim structure information indicating the dependency between the description fragments from the result of the change in the phrase type will be briefly described. As a result of the phrase type change, each token other than JEPSON_CUE, NPT_P, and BP alone become one description fragment. That is, for each token other than JEPSON_CUE, NPT_P, and BP, the delimiter position of the description fragment comes before and after the token. Further, VPR_P and VPK_P are one description fragment including zero or more PPs, VPRs, and VPKs immediately preceding them. That is, for VPR_P and VPK_P, the delimiter position of the description fragment follows. Further, both HP and JEPSON_CUE become one description fragment. That is, for “HP + JEPSSON_CUE”, the delimiter position of the description fragment comes before and after that. The remaining one or more continuous PPs, VPRs, and VPKs also become one description fragment. By using such information indicating the delimitation position of the description fragment corresponding to the token and delimiter clause type, the generation unit 19 can specify the delimitation position of the description fragment. As described above, the result of the phrase type change is shown for each description fragment, and the result of associating the phrase type corresponding to the last clause with the description fragment is as follows.

In the image forming apparatus, JEPSON_CUE
VPK connected to outside
Duct and NPT_P
VPK that discharges air
Air blowing means, NPT_P
VPK to clean air
Cleaning means and NPT_P
COMPOSE_R_CUE
The duct is PP
Internal duct NPT_P
COMPOSE_R_CUE
The cleaning means is installed in the internal duct, VPK_P
FEATURE_K_CUE characterized by
Image forming apparatus. BP

  Further, it is assumed that the dependency relationship between the description fragment to which the token is assigned and the description fragment to which the delimiter clause type is assigned is defined as follows. The description will be divided into the first half and the second half or the whole. In the following description, a description fragment to which a token is assigned or a description fragment to which a phrase type is assigned is represented by the token or the phrase type itself for convenience of explanation. For example, “COMPOSE_K_CUE” strictly means “descriptive fragment to which COMPOSE_K_CUE is assigned”.

(First half)
COMPOSE_K_CUE relates to JEPSON_CUE, FEATURE_CUE, or FEATURE_K_CUE that is closest to and exists after that.
COMPOSE_R_CUE relates to JEPSON_CUE, FEATURE_CUE, or FEATURE_K_CUE that is closest to and exists after that.
FEATURE_CUE relates to JEPSON_CUE.
FEATURE_K_CUE relates to JEPSON_CUE.
NPT_P relates to COMPOSE_K_CUE or COMPOSE_R_CUE that is closest to and exists after that.
VPK relates to the description fragment immediately after.
VPK_P relates to the immediately following description fragment.
The VPR relates to the description fragment immediately after the VPK (or VPK_P) that exists after that and is closest to it.
VPR_P relates to a description fragment immediately after VPK (or VPK_P) that is present and closest to VPR_P.
PP relates to the immediately following description fragment.
(Second half or whole)
JEPSON_CUE relates to BP.
COMPOSE_K_CUE relates to BP, FEATURE_CUE, or FEATURE_K_CUE that is closest to and exists after that.
COMPOSE_R_CUE relates to BP, FEATURE_CUE, or FEATURE_K_CUE that is closest to and exists after that.
FEATURE_CUE relates to BP.
FEATURE_K_CUE relates to BP.
NPT_P relates to COMPOSE_K_CUE or COMPOSE_R_CUE that is closest to and exists after that.
VPK relates to the description fragment immediately after.
VPK_P relates to the immediately following description fragment.
The VPR relates to the description fragment immediately after the VPK (or VPK_P) that exists after that and is closest to it.
VPR_P relates to a description fragment immediately after VPK (or VPK_P) that is present and closest to VPR_P.
PP relates to the immediately following description fragment.

  The generation unit 19 uses dependency information (which may be stored in a recording medium (not shown)) of a description fragment to which a predetermined token is assigned and a description fragment to which a delimiter clause type is assigned. The dependency relationship of each description fragment can be specified. In the case of this example, the destination of each description fragment can be specified. This is the claim structure information shown below. In the claim structure information, “descriptive fragment ID”, “character string of descriptive fragment”, “token or clause type corresponding to descriptive fragment”, “description fragment ID of destination” are displayed in this order. .

[Claim structure information]
1 In the image forming apparatus, JEPSON_CUE 14
2 VPK connected to outside 3
3 Duct and NPT_P 8
4 Discharge air VPK 5
5 blowing means and NPT_P 8
6 VPK to clean air 7
7 Cleaning means and NPT_P 8
8 and COMPOSE_R_CUE 13
9 The duct is PP 10
10 Internal duct NPT_P 11
11 and COMPOSE_R_CUE 13
12 The cleaning means is installed in the internal duct, VPK_P 13
13 FEATURE_K_CUE 14
14 Image forming apparatus. BP

  When the output unit 20 displays the claim structure information so as to indicate the dependency of the description fragment, the display shown in FIG. 8 is obtained. As a method of displaying a plurality of pieces of information having a dependency relationship, for example, software that displays a dependency relationship in natural language processing or software for displaying a claim tree is already known, and a detailed description thereof will be given. Omitted.

  Note that QUIOTE_CUE was not particularly mentioned, but when QUIOTE_CUE exists, the delimiter position of the description fragment exists at the position immediately before it. Further, the description fragment including the QUIOTE_CUE at the head is related to the description fragment of JEPSON_CUE in the first half, and is related to BP in the second half. In this way, even when QUEOTE_CUE is included, it is possible to divide into description fragments and specify the dependency relationship.

  As described above, according to the claim structure information generation device 1 according to the present embodiment, since the clue information on the surface layer and the clue information on the part of speech are used to detect the break position of the description fragment in phrase units, The structural analysis can be appropriately performed even for the claims having a deep structure. In particular, by using the part-of-speech clue information, it is possible to find all the locations that can be the delimiter positions. Note that the break position found in this way may include a portion that should not be a break position. Therefore, the original description fragment can be acquired by deleting such an inappropriate delimiter position by changing the phrase type by the delimiter-phrase type changing unit 18, particularly by deleting the delimiter-phrase type. Become. In the conventional example, the delimiter position of the description fragment is detected in units of morphemes smaller than the clause, so that only a rough analysis can be performed. This is because it is very difficult to perform a detailed analysis because the units handled are fine. On the other hand, in the present embodiment, because it detects the break position of the description fragment in phrase units, it becomes possible to perform a finer analysis than the conventional example, and as a result, a claim with a deeper structure Even now, an appropriate analysis can be performed. In this way, what can be analyzed only to the extent of FIG. 16 in the conventional example can be analyzed finely as shown in FIG. 8 in the present embodiment.

  Further, in the present embodiment, by using the surface layer clue information corresponding to the dependency relationship, the structure of a claim dependent on another claim can be analyzed appropriately. Corresponding claim structure information can be generated.

  In addition, in the claim structure information generation apparatus 1 according to the present embodiment, a structure analysis regarding a parent-child relationship of components may be further performed. In this case, the token that indicates the inclusion relationship of the component and the token that indicates the inclusion relationship of the parent component and the token that indicates the inclusion relationship between the parent component and the token that indicates the inclusion relationship of the component. On the other hand, it is assumed that there is a child delimiter clause type that is a delimiter clause type corresponding to a child component. The token indicating the inclusion relationship of the constituent elements is, for example, the above-mentioned COMPOSE_K_CUE or COMPOSE_R_CUE. Further, the parent break clause type is, for example, the above-described HP or BP, NPH, NPH_P, or the like defined here. The child break clause type is, for example, the above-mentioned NPT_P.

In the phrase type correspondence information, the part-of-speech clue information corresponding to the delimited phrase types NPH and NPH_P is, for example, as follows.
NPH: X + particle "ha"
NPH_P: X + particle "ha" + "symbol-punctuation (", "or", ")"
That is, for example, in the above-described specific example, “the duct” is NPH.

  In such a case, the output unit 20 may display both description fragments so that the parent-child relationship between the description fragment corresponding to the parent break clause type and the description fragment corresponding to the child break clause type can be understood. For example, the output unit 20 may display the description fragment corresponding to the parent delimiter clause type and the description fragment corresponding to the child delimiter clause type by connecting with a line. Note that there may be a description fragment related to the inclusion relationship between the constituent elements between the description fragment corresponding to the parent delimiter clause type and the description fragment corresponding to the child delimiter clause type. That is, the description fragment corresponding to the parent delimiter clause type and the description fragment corresponding to the child delimiter clause type may be connected by a line via the description fragment regarding the inclusion relation of the constituent elements. In addition, the output unit 20 may perform such display as a result. For example, the description fragment corresponding to the parent break clause type and the description fragment corresponding to the child break clause type are directly or indirectly related. If the claim structure information indicating the receiving relationship is generated, it may be displayed as described in the present embodiment.

Here, it demonstrates using a specific example.
(Α) When changing the phrase type, the specific example described above is changed, and the following processing is performed after the change.
(Β) If COMPOSE_K_CUE or COMPOSE_R_CUE does not exist after NPH, NPH is changed to PP.
(Γ) If NPT_P does not exist between NPH and COMPOSE_K_CUE or COMPOSE_R_CUE that is present and closest to NPH, change NPH to PP.
(Δ) If VPR_P exists between NPH and NPT_P that is present and closest to NPH, NPH is changed to PP.
(Ε) NPH_P is also replaced with NPH_P and the processes (β) to (δ) are performed.

By performing the processes (α) to (δ), the result of the phrase type change is as follows.
[Result of phrase type change]
Image forming device HP
In JEPSON_CUE
PP outside
Connected VPK
Duct and NPT_P
PP for air
VPK to discharge
Air blowing means, NPT_P
PP for air
Cleaning PP
VPK
Cleaning means and NPT_P
COMPOSE_R_CUE
The duct is NPH
Internal duct NPT_P
COMPOSE_R_CUE
The cleaning means is PP
PP inside the internal duct
VPK_P installed
FEATURE_K_CUE characterized by
Image forming apparatus. BP

  Next, it is assumed that NPH and NPH_P also become one description fragment when specifying the delimiter position of the description fragment. That is, for NPH and NPH_P, description fragment delimiters come before and after the NPH and NPH_P. Further, the relationship of the next NPH and NPH_P is added to the predetermined dependency relationship, and the relationship of the COMPOSE_K_CUE or COMPOSE_R_CUE is changed as follows.

(First half)
(Addition) NPH is related to JEPSON_CUE, FEATURE_CUE, or FEATURE_K_CUE that is closest to and exists after that.
(Addition) NPH_P relates to JEPSON_CUE, FEATURE_CUE, or FEATURE_K_CUE that is present closest to the NPH_P.
(Change) If COMPE_K_CUE has NPH or NPH_P before it, it is related to NPH or NPH_P that is closest to that COMPOSE_K_CUE, and NPH and NPH_P exist before that COMPOSE_K_CUE If not, it relates to JEPSON_CUE, FEATURE_CUE, or FEATURE_K_CUE that is closest to and exists after that COMPOSE_K_CUE.
(Change) When COMPOSE_R_CUE has NPH or NPH_P before it, it is related to NPH or NPH_P that is closest to that COMPOSE_R_CUE, and has NPH and NPH_P before that COMPOSE_R_CUE. If not, it is related to JEPSON_CUE, FEATURE_CUE, or FEATURE_K_CUE which is present after that COMPOSE_R_CUE and is closest thereto.
(Second half or whole)
(Addition) NPH is related to BP, FEATURE_CUE, or FEATURE_K_CUE that is closest to and exists after that.
(Addition) NPH_P relates to BP, FEATURE_CUE, or FEATURE_K_CUE that is present and closest to it.
(Change) If COMPE_K_CUE has NPH or NPH_P before it, it is related to NPH or NPH_P that is closest to that COMPOSE_K_CUE, and NPH and NPH_P exist before that COMPOSE_K_CUE If not, it relates to BP, FEATURE_CUE, or FEATURE_K_CUE that is closest to and exists after that COMPOSE_K_CUE.
(Change) When COMPOSE_R_CUE has NPH or NPH_P before it, it is related to NPH or NPH_P that is closest to that COMPOSE_R_CUE, and has NPH and NPH_P before that COMPOSE_R_CUE. If not, it is related to BP, FEATURE_CUE, or FEATURE_K_CUE which is present after that COMPOSE_R_CUE and is closest thereto.

When the claim structure information is generated according to such a dependency relationship, it is as follows.
[Claim structure information]
1 In the image forming apparatus, JEPSON_CUE 14
2 VPK connected to outside 3
3 Duct and NPT_P 8
4 Discharge air VPK 5
5 blowing means and NPT_P 8
6 VPK to clean air 7
7 Cleaning means and NPT_P 8
8 and COMPOSE_R_CUE 13
9 The duct is NPH 13
10 Internal duct NPT_P 11
11 and COMPOSE_R_CUE 9
12 The cleaning means is installed in the internal duct, VPK_P 13
13 FEATURE_K_CUE 14
14 Image forming apparatus. BP

  Further, when the output unit 20 displays the claim structure information, for example, as shown in FIG. In FIG. 9, description fragments of components having a parent-child relationship are connected via “comprises”, “characteristically”, or “has”. For example, “image forming device” is connected to “image forming apparatus” via “comprising” and the like, with “duct and”, “air blowing device”, and “cleaning device”. It can be seen that the "apparatus" has "duct," "blower means," and "cleaning means" as components. In addition, it is understood that “the duct” has “inner duct” as a constituent element because “the duct” is connected to “the duct” via “has”.

  Note that the output unit 20 may display a plurality of description fragments related to the constituent elements in common with the same stem. The stem of a description fragment is a punctuation mark or particle at the end, or a prefix indicating an existing one such as “above”, “above”, “this”, “this”, “that”, etc. from the description fragment. It is the removed part. Therefore, “the duct” and “the duct” have the common word “duct”. In the case of commonization, other description fragments may be integrated into the first description fragment among a plurality of description fragments with a common stem, or other description fragments may be integrated into the description fragment related to COMPOSE_K_CUE or COMPOSE_R_CUE May be. As a result, for example, the display shown in FIG. 10 is performed. Here, the case where the output unit 20 performs such output has been described, but the generation unit 19 may generate such claim structure information. For example, in the case of the above-described example, in the claim structure information, the description fragments related to the constituent elements are NPT_P, NPH, and NPH_P, so it is determined whether or not their stems are common. In this case, it is determined that the stems of “duct and” and “the duct is” are common. As a result, the description fragment “having a description fragment“ having a duct ”related to the second and subsequent description fragments is changed to the first description fragment“ duct ”, and the second and subsequent description fragments. It is only necessary to delete “the duct”. As a result, the claim structure information is as follows.

[Claim structure information]
1 In the image forming apparatus, JEPSON_CUE 14
2 VPK connected to outside 3
3 Duct and NPT_P 8
4 Discharge air VPK 5
5 blowing means and NPT_P 8
6 VPK to clean air 7
7 Cleaning means and NPT_P 8
8 and COMPOSE_R_CUE 13
10 Internal duct NPT_P 11
11 and COMPOSE_R_CUE 3
12 The cleaning means is installed in the internal duct, VPK_P 13
13 FEATURE_K_CUE 14
14 Image forming apparatus. BP

  In this example, a case has been described in which description fragments relating to constituent elements that share a common word stem are collected in one claim, but this may be performed in a dependent claim. For example, in a claim dependent on the claim of the above example, when there is a description that "... the air blowing means includes a blade and a motor that rotates the blade, ..." The description fragment “blade” and “motor” may be displayed with respect to the description fragment “with air blowing means” in the section.

  In the claim structure information generating apparatus 1 according to the present embodiment, the parent-child relationship between the constituent elements may be acquired using a parent-delimited clause type or a child-delimited clause type. In that case, the claim structure information generation device 1 may further include a parent-child relationship acquisition unit 21 and a hierarchy level acquisition unit 22, as shown in FIG.

  The parent-child relationship acquisition unit 21 acquires the parent-child relationship of the constituent elements using the patent claim information to which the parent-delimited clause type and the child-delimited clause type are assigned. Specifically, the clause component corresponding to the parent delimiter clause type for a token indicating the inclusion relationship of a certain component is the parent component, and the clause component corresponding to the child delimiter clause type for the same token Is a child component. Therefore, the parent-child relationship of the component can be acquired by interposing a token indicating the inclusion relationship of the component. The parent-child relationship may be, for example, information indicating that the child corresponding to the parent image forming apparatus is a duct, a blowing unit, or a cleaning unit. As in this example, for convenience of explanation, an apparatus or the like is also referred to as a component. Further, the method for specifying the constituent element from the description fragment is the same as the method for specifying the word stem, and the description thereof is omitted. Information indicating the parent-child relationship may be stored in a recording medium (not shown).

  The hierarchy level acquisition unit 22 acquires the hierarchy level of the component using the parent-child relationship of the component acquired by the parent-child relationship acquisition unit 21. Although a method for acquiring a hierarchical level from a parent-child relationship is known, an example thereof will be briefly described. Here, it is assumed that the highest hierarchical level (hierarchical level for which no parent exists) is “1”, and the value of the hierarchical level increases as the hierarchy becomes deeper. First, all components are acquired from the parent-child relationship, and unique processing is performed. Then, a hierarchy level 1 is assigned to a component that does not have a parent among the components. Hierarchy level 2 is assigned to all the components having the hierarchy level 1 component as a parent. In this way, the process of assigning the hierarchy level (N + 1) to all the constituent elements having the constituent element of the hierarchical level N as a parent is continued until there is no constituent element to which no hierarchical level is assigned. As a result, it becomes possible to know the hierarchical level of all the constituent elements. Information indicating the hierarchical level for each component may be stored in a recording medium (not shown).

  In addition, the output unit 20 may display description fragments corresponding to the parent break phrase type and the child break phrase type according to the hierarchy level acquired by the hierarchy level acquisition unit 22. “Displaying a description fragment according to the hierarchy level” may be, for example, displaying a numerical value indicating the hierarchy level in association with the description fragment, or a display position ( For example, the description fragment may be displayed so that the higher the hierarchy level, the left (or upper) the display fragment may be displayed, and the lower the hierarchy level, the right (or lower) may be displayed.

Here, it demonstrates using a specific example.
The process until the claim structure information is generated is the same as the process described above, and the description thereof is omitted. The claim structure information is as follows.

[Claim structure information]
1 In the image forming apparatus, JEPSON_CUE 14
2 VPK connected to outside 3
3 Duct and NPT_P 8
4 Discharge air VPK 5
5 blowing means and NPT_P 8
6 VPK to clean air 7
7 Cleaning means and NPT_P 8
8 and COMPOSE_R_CUE 13
9 The duct is NPH 13
10 Internal duct NPT_P 11
11 and COMPOSE_R_CUE 9
12 The cleaning means is installed in the internal duct, VPK_P 13
13 FEATURE_K_CUE 14
14 Image forming apparatus. BP

  The parent-child relationship acquisition unit 21 acquires a parent-child relationship for each COMPOSE_K_CUE or COMPOSE_R_CUE. If NPH or NPH_P exists before COMPOSE_K_CUE or COMPOSE_R_CUE, the nearest NPH or NPH_P stem is obtained as the parent component for that COMPOSE_K_CUE or COMPOSE_R_CUE. In addition, the NPT_P stem of the COMPOSE_K_CUE or COMPOSE_R_CUE is acquired as a child component. In this way, the parent-child relationship can be acquired.

  Further, when NPH or NPH_P does not exist before COMPOSE_K_CUE or COMPOSE_R_CUE, the parent-child relationship acquisition unit 21 uses the relationship of the COMPOSE_K_CUE or COMPOSE_R_CUE as the parent component for the COMPOSE_K_CUE or COMPOSE_R_CUE The description fragment including the closest HP (that is, the description fragment of JEPSON_CUE including the HP) or the stem of the BP is obtained. Further, the parent-child relationship acquisition unit 21 acquires the NPT_P stem associated with the COMPOSE_K_CUE or COMPOSE_R_CUE as a child component. In this way, the parent-child relationship can be acquired.

  In the case of the above-described claim structure information, the parent-child relationship acquisition unit 21 includes, for example, a description component “8, COMPOSE_R_CUE 13”, a parent component “image forming apparatus”, and a child component “ “Duct”, “Blowing means”, “Cleaning means” can be acquired. Further, the parent-child relationship acquisition unit 21 can acquire, for example, the parent component “duct” and the child component “internal duct” for the description fragment “11, COMPOSE_R_CUE 9”. . Therefore, the information indicating the acquired parent-child relationship is as shown in FIG.

The hierarchical level acquisition unit 22 acquires the hierarchical level of each component as follows from the parent-child relationship shown in FIG. 12, for example, by the method described above.
Component Hierarchy level Image forming device 1
Duct 2
Blowing means 2
Cleaning means 2
Internal duct 3

  When displaying the description fragment using the correspondence between the component and the hierarchy level, the output unit 20 corresponds to the component if the stem of the description fragment matches any component. The numerical value of the hierarchy level to be displayed is displayed at the position of the description fragment. As a result, the display of FIG. 13 is made. In this way, the hierarchical level of each component can also be displayed, and the user who sees it can easily understand the hierarchical level of the component. In this example, the hierarchical level is such that the numerical value increases as the hierarchical level increases. That is, a hierarchical level in which the numerical value decreases as the hierarchical level increases.

  In this embodiment, when displaying a description fragment including a constituent element (for example, a description fragment to which BP, NPT_P, NPH, NPH_P, etc.) is displayed, only the word stem of the description fragment is displayed. May be. For example, the stem “image forming apparatus” may be displayed instead of the description fragment “image forming apparatus”, and the stem “duct” may be displayed instead of the description fragment “the duct”. .

  Further, in the present embodiment, a case has been described in which the phrase type is changed, such as deletion of the delimited phrase type, but this need not be the case. When the phrase type is not changed, the claim structure information generation device 1 may not include the delimiter phrase type change unit 18. Even if the clause type is not changed, rules indicating the structure of the claimed claim (for example, rules described in a context-free grammar, tokens or delimited clause types are added) It is also possible to obtain the same result as when the phrase type is changed by making the dependency relationship of the description fragment appropriate.

Needless to say, the specific rules, conditions, information, and the like used in the above embodiments are not limited thereto. As long as the object can be realized, they may be appropriately changed.
Further, in the above embodiment, the case where the claim structure information generation device 1 is stand-alone has been described. However, the claim structure information generation device 1 may be a stand-alone device, or a server device in a server / client system. It may be. In the latter case, the output unit may output information via a communication line.

  In the above embodiment, each process or each function may be realized by centralized processing by a single device or a single system, or may be distributedly processed by a plurality of devices or a plurality of systems. It may be realized by doing.

  In the above embodiment, information related to processing executed by each component, for example, information received, acquired, selected, generated, transmitted, or received by each component In addition, information such as threshold values, mathematical formulas, addresses, etc. used by each component in processing is retained temporarily or over a long period of time on a recording medium (not shown) even when not explicitly stated in the above description. It may be. Further, the storage of information in the recording medium (not shown) may be performed by each component or a storage unit (not shown). Further, reading of information from the recording medium (not shown) may be performed by each component or a reading unit (not shown).

  In the above embodiment, when information used by each component, for example, information such as a threshold value, an address, and various setting values used by each component may be changed by the user Even if it is not specified in the above description, the user may be able to change the information as appropriate, or it may not be. If the information can be changed by the user, the change is realized by, for example, a not-shown receiving unit that receives a change instruction from the user and a changing unit (not shown) that changes the information in accordance with the change instruction. May be. The change instruction received by the receiving unit (not shown) may be received from an input device, information received via a communication line, or information read from a predetermined recording medium, for example. .

  In the above embodiment, when two or more constituent elements included in the claim structure information generating apparatus 1 have a communication device, an input device, etc., the two or more constituent elements have a physically single device. Or you may have separate devices.

  In the above embodiment, each component may be configured by dedicated hardware, or a component that can be realized by software may be realized by executing a program. For example, each component can be realized by a program execution unit such as a CPU reading and executing a software program recorded on a recording medium such as a hard disk or a semiconductor memory. In addition, the software which implement | achieves the claim structure information generation apparatus 1 in the said embodiment is the following programs. In other words, this program is a program for causing a computer to function as a claim structure information generation device that generates claim structure information indicating a description fragment in which the text of a claim is delimited and the structure of the description fragment. A morpheme analyzer for morphological analysis of the claim information stored in the storage unit for storing the claim information which is the text of the claim, and a phrase specification for specifying the break position of the clause of the claim information Part, claim information, and a surface layer delimiter in which two or more surface layer delimiter information are stored that include descriptive fragment delimiters and surface layer clue information indicating the relationship between the descriptive fragments and a token corresponding to the surface layer clue information Token grant that grants a token corresponding to the surface clue information at a position corresponding to the surface clue information stored in the information storage unit In the claim information, a phrase type correspondence information storage unit that stores two or more phrase type correspondence information having clue information on the part of speech of the clause that is a delimiter of the description fragment and a segment phrase type corresponding to the clue information on the part of speech In the clause corresponding to the part-of-speech clue information stored in the above, a separator-phrase type assigning unit for assigning a clause-type corresponding to the part-of-speech clue information, a token and a clause-type specified in the claim information are used. , In accordance with a rule indicating the structure of a patent claim defined in advance, a generation unit for generating claim structure information indicating the structure of a description fragment of claim claim information, and outputting the claim structure information generated by the generation unit It is for functioning as an output unit.

  In the program, the functions realized by the program do not include functions that can be realized only by hardware. For example, a function that can be realized only by hardware such as a modem or an interface card in an output unit that outputs information is not included in at least the function realized by the program.

  Further, this program may be executed by being downloaded from a server or the like, and a program recorded on a predetermined recording medium (for example, an optical disk such as a CD-ROM, a magnetic disk, a semiconductor memory, or the like) is read out. May be executed by Further, this program may be used as a program constituting a program product.

  Further, the computer that executes this program may be singular or plural. That is, centralized processing may be performed, or distributed processing may be performed.

  FIG. 14 is a schematic diagram showing an example of an external appearance of a computer that executes the program and realizes the claim structure information generation device 1 according to the embodiment. The above-described embodiment can be realized by computer hardware and a computer program executed on the computer hardware.

  14, a computer system 900 includes a computer 901 including a CD-ROM (Compact Disk Read Only Memory) drive 905 and an FD (Floppy (registered trademark) Disk) drive 906, a keyboard 902, a mouse 903, a monitor 904, and the like. Is provided.

  FIG. 15 is a diagram showing an internal configuration of the computer system 900. In FIG. 15, in addition to the CD-ROM drive 905 and the FD drive 906, a computer 901 is connected to an MPU (Micro Processing Unit) 911, a ROM 912 for storing a program such as a bootup program, and the MPU 911. A RAM (Random Access Memory) 913 that temporarily stores program instructions and provides a temporary storage space, a hard disk 914 that stores application programs, system programs, and data, and an MPU 911 and a ROM 912 are interconnected. And a bus 915. The computer 901 may include a network card (not shown) that provides connection to the LAN.

  A program for causing the computer system 900 to execute the function of the claimed structure information generating apparatus 1 according to the above-described embodiment is stored in the CD-ROM 921 or FD 922 and inserted into the CD-ROM drive 905 or FD drive 906. It may be transferred to the hard disk 914. Instead, the program may be transmitted to the computer 901 via a network (not shown) and stored in the hard disk 914. The program is loaded into the RAM 913 when executed. The program may be loaded directly from the CD-ROM 921, the FD 922, or the network.

  The program does not necessarily include an operating system (OS), a third party program, or the like that causes the computer 901 to execute the functions of the claim structure information generation device 1 according to the above-described embodiment. The program may include only a part of an instruction that calls an appropriate function (module) in a controlled manner and obtains a desired result. How the computer system 900 operates is well known and will not be described in detail.

  Further, the present invention is not limited to the above-described embodiment, and various modifications are possible, and it goes without saying that these are also included in the scope of the present invention.

  As described above, according to the claim structure information generating apparatus and the like according to the present invention, an effect that a structure analysis can be performed on a claim having a deeper structure can be obtained. Useful.

DESCRIPTION OF SYMBOLS 1 Claim structure information production | generation apparatus 11 Memory | storage part 12 Morphological analysis part 13 Clause specific part 14 Surface layer division | segmentation information storage part 15 Token provision part 16 Phrase type corresponding | compatible information storage part 17 Separation phrase type provision part 18 Separation phrase type change part 19 Generation part 20 Output unit 21 Parent-child relationship acquisition unit 22 Hierarchy level acquisition unit

Claims (10)

A claim structure information generating device for generating claim structure information indicating a description fragment in which a text of a claim is delimited and a structure of the description fragment,
A storage unit for storing claim information which is a text of a claim;
A morpheme analysis unit for morphological analysis of the claim information;
A phrase specifying unit for specifying a section position of the clause of the claim information;
A surface layer delimiter information storage unit for storing two or more surface layer delimiter information including delimiters of description fragments and surface layer clue information indicating a relationship between the description fragments, and a token corresponding to the surface layer clue information;
In the claim information, a token granting unit for granting a token corresponding to the surface clue information at a position corresponding to the surface clue information,
A phrase type correspondence information storage unit for storing two or more phrase type correspondence information having clue information on the part of speech of a clause that is a delimiter of a description fragment and a segment clause type corresponding to the clue information on the part of speech;
In the claim information, a delimiter clause type giving unit that gives a delimiter clause type corresponding to the clue information of the part of speech to a clause corresponding to the clue information of the part of speech,
A claim structure indicating a structure of a description fragment of the claim information according to a rule indicating a structure of the claim, which is defined in advance, using a token and a delimiter clause type given to the claim information. A generator for generating information;
An output unit that outputs claim structure information generated by the generation unit , and
The generator is
Using a parser generated using a grammar file having rules relating to tokens and delimited clause types described in a context-free grammar and an action for assigning a structure corresponding to the rules to the claim information, A claim structure information generation device that generates claim structure information corresponding to the claim information. A delimiter clause type changing unit that deletes delimiter clause types existing at positions that are not delimiters of the descriptive fragment according to the sequence of descriptive fragments,
The generation unit generates claim structure information after the delimiter clause type change unit deletes the delimiter clause type ,
The delimiter clause type change part is:
Termination delimiter between a delimiter clause type that is the delimiter clause type at the end of parallel description fragments and a continuation delimiter clause type that is a delimiter clause type that can be continued by parallel descriptive fragments before the terminal delimiter clause type The claim structure information generating device according to claim 1, wherein when a clause type exists, the terminal break clause type existing between both break clause types is deleted . A delimiter clause type changing unit that deletes delimiter clause types existing at positions that are not delimiters of the descriptive fragment according to the sequence of descriptive fragments,
The generation unit generates claim structure information after the delimiter clause type change unit deletes the delimiter clause type,
The delimiter clause type change part is:
The claim structure information generation device according to claim 1 , wherein when there is a line break and there is a line break type related to a verb not followed by a line break, the line break type is deleted. A description in the context-free grammar of the structure of the claim is written in the direction from the end to the beginning of the claim,
The generator is
The claim structure information generation according to any one of claims 1 to 3 , wherein the claim structure information is generated by taking in the claim information to which a token and a delimiter clause type are attached in a direction from a terminal to a head. apparatus. The generator is
Claim structure information described using a markup language is generated by adding a mark indicating a structure corresponding to a rule included in the grammar file to the claim information. 5. The structure information generating device according to claim 4 . The output unit is
The claim structure information generating apparatus according to any one of claims 1 to 5 , wherein the claim structure information generated by the generation unit is displayed so that a relationship between description fragments can be understood. Tokens exist according to the inclusion relationship of components,
For delimiter clause types,
For tokens that indicate the inclusion relationship of components, a parent delimiter clause type that is a delimiter clause type according to the parent component,
There is a child delimiter clause type, which is a delimiter clause type corresponding to the child component, for the token indicating the inclusion relationship of the component,
7. The structure information generation according to claim 6 , wherein the output unit displays both description fragments so that a parent-child relationship between a description fragment corresponding to a parent-delimited clause type and a description fragment corresponding to a child-delimited clause type can be understood. apparatus. A parent-child relationship acquisition unit for acquiring a parent-child relationship of the constituent elements using the claim information to which a parent-delimited clause type and a child-delimited clause type are attached;
A hierarchy level acquisition unit that acquires the hierarchy level of the component using the parent-child relationship of the component acquired by the parent-child relationship acquisition unit, and
8. The structure information generating apparatus according to claim 7 , wherein the output unit displays a description fragment corresponding to a parent delimiter clause type and a child delimiter clause type according to the hierarchy level acquired by the hierarchy level acquisition unit. A storage unit for storing claim information, which is a text of a claim, a morpheme analysis unit, a phrase specifying unit, and a surface layer indicating a relationship between the description fragments in the claim text Clue information and a part-of-speech clue information of a clause serving as a delimiter for a description fragment, a surface layer delimiter information storage unit storing two or more surface layer delimiter information having tokens corresponding to the surface layer clue information And a phrase type correspondence information storage unit that stores two or more phrase type correspondence information having a segmentation phrase type corresponding to the clue information of the part of speech, a segmentation phrase type giving unit, a generation unit, and an output unit Claim structure information generating method for generating claim structure information indicating a description fragment and a structure of the description fragment, which are processed by
The morpheme analysis unit performs a morpheme analysis step for morphological analysis of the claim information;
The phrase specifying unit for specifying the position where the clause of the claim information is separated;
In the claim information, the token granting unit grants a token corresponding to the surface clue information at a position corresponding to the surface clue information; and
In the claim information, the demarcation phrase type giving unit assigns a demarcation phrase type corresponding to the part-of-speech clue information to a phrase corresponding to the part-of-speech clue information; and
Structure of description fragment of claim information according to a rule indicating a structure of claim according to a predetermined rule, wherein the generation unit uses a token and a delimiter clause type given to the claim information. Generating a claim structure information indicating:
An output step for outputting the claim structure information generated in the generation step ;
In the generating step,
Using a parser generated using a grammar file having rules relating to tokens and delimited clause types described in a context-free grammar and an action for assigning a structure corresponding to the rules to the claim information, A claim structure information generation method for generating claim structure information corresponding to the claim information. Computer
A program for functioning as a claim structure information generation device for generating claim structure information indicating a description fragment in which a text of a claim is delimited and a structure of the description fragment,
A morpheme analysis unit that performs morphological analysis on claim information stored in a storage unit that stores claim information that is a text of a claim,
A phrase specifying unit for specifying a section position of the clause of the claim information;
In the claim information, a surface layer delimiter in which two or more surface layer delimiter information including a surface layer clue information indicating a relationship between the description fragments and a token corresponding to the surface layer clue information is stored. A token grant unit that grants a token corresponding to the surface clue information at a position corresponding to the surface clue information stored in the information storage unit;
In the claim information, the phrase type correspondence information storage in which two or more phrase type correspondence information having clue information on the part of speech of the clause serving as a delimiter of the description fragment and a break phrase type corresponding to the clue information on the part of speech are stored. A delimiter-phrase type assigning unit that adds a delimiter-phrase type corresponding to the clue information of the part of speech to the clause corresponding to the clue information of the part of speech stored in the part,
A claim structure indicating a structure of a description fragment of the claim information according to a rule indicating a structure of the claim, which is defined in advance, using a token and a delimiter clause type given to the claim information. A generator for generating information,
Function as an output unit that outputs claim structure information generated by the generation unit ;
The generator is
Using a parser generated using a grammar file having rules relating to tokens and delimited clause types described in a context-free grammar and an action for assigning a structure corresponding to the rules to the claim information, A program for generating claim structure information corresponding to the claim information .

Images