JP7203554B2 - Information processing equipment - Google Patents

Description

The present invention relates to an information processing device, an information processing method, and a program.

2. Description of the Related Art There has long been a need to easily find a document containing desired information or a reference document from a large number of documents (for example, papers, technical data, patent documents, etc.). Conventionally, as an approach to such needs, a method of extracting a plurality of documents that match a search query and displaying a list in descending order of degree of matching has been the mainstream method. However, with such a method, even if a user sees a document list output as a search result, the user cannot grasp the relevance or similarity between the extracted documents. It was difficult. On the other hand, Non-Patent Document 1 proposes an attempt to intuitively express the relationship between documents by displaying a plurality of words picked up from extracted documents in a tree structure.

Scott Spangler et.al., "Automated Hypothesis Generation Based on Mining Scientific Literature"

However, when the inventors of the present invention have verified that the expression by tree structure is very useful, the method of Non-Patent Document 1 cannot appropriately express the relationship between words and the relationship/similarity between documents. There are many problems, and we have come to recognize the problem that further improvement is necessary for practical use.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned circumstances. The purpose is to provide technology that can support

According to one aspect of the present invention, a search unit extracts a plurality of documents matching a search condition as extracted documents from a group of documents accumulated in a database; an analysis unit for extracting a plurality of character strings from a document as an extracted character string; a document feature calculation unit for calculating a document feature score obtained by digitizing a feature of each of the plurality of extracted documents; a character string feature calculator for obtaining, for each of the extracted character strings, a character string feature quantity representing the features of the extracted character string from document feature scores of one or more extracted documents including the extracted character string; and the plurality of extracted character strings. is associated with a node, and each node is arranged based on the difference in character string feature amounts between extracted character strings.

A "character string" may be a "word", or may be a "compound word" or a "phrase" composed of a plurality of words. A “character string feature amount” may be an index (scalar) consisting of a single value or an index (vector) consisting of a set of a plurality of values. In the case of a scalar, the "character string feature amount difference" is, for example, the difference between the character string feature amounts of two character strings or its absolute value. In the case of vectors, the "character string feature amount difference" can be calculated from, for example, cosine similarity or Euclidean distance between two vectors.

In the tree structure of the present invention described above, the arrangement of each node is determined based on the difference in the character string feature quantity representing the characteristics of the character strings. You can easily understand the tendency of character strings included in the results (multiple extracted documents).

In addition, the feature of the extracted character string is expressed using the document feature score, which is the feature of the document (text) using the extracted character string, rather than the extracted character string itself. Therefore, the arrangement and connection relationship of each node in the tree structure well reflect the relationship and similarity between documents. Therefore, by using the tree structure of the present invention, it is possible to appropriately and intuitively express the relationship/similarity between documents and the relationship between words appearing in a plurality of documents. It becomes possible to support work.

A scalar character string feature may be calculated as follows. For example, the character string feature calculation unit may use a statistical value calculated from the document feature scores of one or more extracted documents including the character string as the character string feature amount of the character string. A "statistical value" is, for example, the total value, average value, or standard deviation of the document feature scores of one or more extracted documents containing the relevant character string.

A vector character string feature may be calculated as follows. For example, the character string feature calculation unit may use a vector having, as elements, values based on document feature scores of one or more extracted documents including the character string, as the character string feature quantity of the character string. The "value based on the document feature score" is, for example, a value obtained by multiplying the document feature score of the extracted document by the appearance frequency of the character string in the extracted document.

The document feature calculation unit includes a classifier that classifies an input document into n classes (n is an integer equal to or greater than 2). It may be a document feature score. The “classifier” may be, for example, a classifier generated by machine learning using a plurality of documents as teacher data, or a classifier theoretically created from rules and models.

In addition to the method of obtaining the character string feature amount from the document feature score, a method of obtaining the character string feature amount directly from the character string can also be adopted. For example, an information processing apparatus includes a search unit that extracts a plurality of documents matching a search condition as extracted documents from a group of documents accumulated in a database; an analysis unit for extracting a plurality of character strings as extracted character strings from the input It consists of a character string classifier that classifies character strings into n classes (n is an integer equal to or greater than 2). a character string feature calculation unit as a column feature amount, each of the plurality of extracted character strings is associated with a node, and each node is arranged based on a difference in character string feature amounts between the extracted character strings, and an output unit for outputting a tree structure. This "character string classifier" may be, for example, a classifier generated by machine learning using a plurality of character strings as teacher data, or a classifier theoretically created from rules and models.

When the search condition includes a keyword, the search unit obtains synonyms of the keyword from a synonym dictionary that stores synonyms having the same meaning, and includes at least one of the keyword and the synonyms. A document may be extracted as the extracted document. According to such a technology, by searching synonyms of keywords as well, it is possible to suppress search omissions even when there are fluctuations in the notation of character strings between documents, and good search results that reflect the user's search intentions can be obtained. can be obtained.

A base point determination unit that calculates an average of character string feature amounts among the plurality of character strings and determines a character string having a character string feature amount closest to the average among the plurality of character strings as a base point of the tree structure. may have According to this technique, the average character string (when viewed in terms of character string feature values) is selected as the base point from among multiple character strings, so that branches (edges) branching from that base point are well-balanced. It can be expected that the nodes will be connected, and as a result, a tree structure with little bias will be obtained.

A base point changing unit may be further provided for changing the base point of the tree structure to a character string specified by the user. According to such technology, the base point of the tree structure can be changed as intended by the user. For example, by observing how the tree structure changes while changing the base character string, and comparing the generated tree structures, you can find a tree structure that is easy to obtain the analysis results you expect. There is a possibility that you will be able to find headlines or discover new relationships between character strings or between documents.

The base point changing unit causes the user to specify any node of the output tree structure, and changes the base point of the current tree structure to a character string associated with the node specified by the user. good too. With such a user interface, the user can easily change the base point, which is highly convenient.

The output unit outputs a node corresponding to a predetermined character string, and/or a character string associated with an adjacent node corresponding to the predetermined character string and the predetermined character string. and/or a node that satisfies a predetermined condition with an adjacent node may be output in a manner different from other nodes. "Different mode" may be any display effect, such as changing color, changing design or typeface, changing character size, adding movement such as blinking or animation. According to such a technique, for example, only a character string that the user is paying attention to or a character string that is highly similar or related to the character string can be highlighted.

Each of the document groups stored in the database includes reference information indicating other documents to be referred to, and the search unit searches for documents with a small number of references from other documents based on the reference information. You may preferentially extract. A document with a small number of references is highly likely to be a document containing information that is not well known. Therefore, such a technique has the advantage of preferentially extracting documents from which the user is likely to obtain new knowledge.

Each of the document groups stored in the database includes reference information indicating other documents to be referred to, and the search unit searches for a document that is referred to by other documents with a large number of references based on the reference information. You may preferentially extract. A document with a large number of references is highly likely to be a document of high importance. Therefore, such a technique has the advantage of preferentially extracting documents from which important information is highly likely to be obtained.

The analysis unit may extract, as one character string, consecutive words having a predetermined number of words or less. Also, the user may be allowed to change the setting of the "predetermined number of words".

The output unit may output information of an extracted document including a character string associated with the specified node, when any node of the tree structure is specified by the user. At this time, a text part including the "character string associated with the designated node" in the extracted document may be output as the "extracted document information". Such a user interface facilitates access to documents from (the character strings displayed in) the tree structure, improving convenience.

When any edge of the tree structure is specified by the user, the output unit outputs extracted document information including two character strings respectively associated with two nodes connected by the specified edge. may be output. At this time, as the "extracted document information", a text portion containing each of "two character strings respectively associated with two nodes connected by a specified edge" in the extracted document may be output. good. Such a user interface facilitates access to a document highly related to two adjacent character strings in the tree structure, improving convenience.

In addition, the present invention provides an information processing method including at least part of the functions or processes described above, a program for causing a computer to execute each step of the information processing method, or a storage medium that non-temporarily stores the program. It can also be taken as The present invention also provides a tree structure generation device and a tree structure generation method for generating the tree structure described above, a tree structure output device and tree structure output method for outputting or displaying the tree structure described above, and a tree structure output method for analyzing a plurality of documents. Document analysis device and document analysis method, character string analysis device and character string analysis method for analyzing multiple character strings contained in a document, information search support device and information search support method for supporting information search by users, etc. can also

For a plurality of documents, it is possible to appropriately and intuitively express the relationship/similarity between documents and the relationship between words appearing in the documents, thereby supporting the information search work by the user.

FIG. 1 is a diagram showing an example of the configuration of an information processing apparatus. FIG. 2 is a diagram illustrating an example of functional blocks of the information processing apparatus according to the first embodiment; FIG. 3 is a diagram showing an example of the configuration of a synonym dictionary. FIG. 4 is a diagram illustrating an example of a morphological analysis result by a morphological analysis unit; FIG. 5 is a diagram showing an example of a document vector. FIG. 6 is a diagram showing an example of word vectors. FIG. 7 is a diagram showing an example of a stopword dictionary. FIG. 8 is a diagram showing an example of a classification degree vector. FIG. 9 is a diagram showing an example of a state in which a node of a word with the highest classification degree and a node of a word with the lowest classification degree are arranged under the "empty" base node. FIG. 10 is an example of a state in which the most similar nodes have been added. FIG. 11 is a diagram showing an example of an attribute data list. FIG. 12 is a diagram showing an example of a specific feature data list. FIG. 13 is a diagram showing an example of a proximity rule list. FIG. 14 is a diagram exemplifying a case where the node display mode is changed based on the rules registered in the proximity rule list. FIG. 15 is a diagram showing an example of a display target data list. FIG. 16 is a display example of search results by the detailed data search unit. 17 is a first diagram illustrating an example of a processing flow according to the embodiment; FIG. FIG. 18 is a second diagram illustrating an example of the processing flow according to the embodiment; FIG. 19 is a third diagram illustrating an example of the processing flow according to the embodiment; FIG. 20 is a fourth diagram illustrating an example of the processing flow according to the embodiment; FIG. 21 is a diagram showing an example of a document vector according to the first modification, in which synonyms are collectively generated. FIG. 22 is a diagram illustrating an example of functional blocks of an information processing apparatus according to the second embodiment;

Hereinafter, an information processing apparatus, an information processing method, and a program according to embodiments of the present invention will be described with reference to the drawings. The information processing apparatus according to this embodiment extracts a plurality of documents that match a search condition from a large number of documents stored in a database, and plots relationships between character strings appearing in the extracted documents in a tree-structured graph. format. In the following, two embodiments in which methods of obtaining character string feature amounts indicating character string characteristics are different will be exemplified. The first embodiment discloses a method of indirectly obtaining a character string feature amount using a document feature amount (document feature score). Disclosed is a method for obtaining a character string feature quantity in a systematic manner. However, the configurations of the embodiments shown below are examples of the configurations of the present invention, and the present invention is not limited to the configurations of the following embodiments.

<First Embodiment>
FIG. 1 is a diagram showing an example of the configuration of an information processing apparatus 100 according to the first embodiment. FIG. 1 also illustrates a display 210, a keyboard 220 and a mouse 230 connected to the information processing apparatus 100. FIG. The information processing device 100 has a central processing unit (CPU
) 101 , a main storage unit 102 , an auxiliary storage unit 103 , a communication unit 104 , and an input/output interface (denoted as an input/output IF in the drawing) 105 . CPU 101, main storage unit 102, auxiliary storage unit 103, communication unit 104 and input/output interface 105 are interconnected by connection bus B1.

The CPU 101 is also called a microprocessor unit (MPU) or processor. The CPU 101 is not limited to a single processor, and may have a multiprocessor configuration. Also, a single CPU 101 connected with a single socket may have a multi-core configuration. At least part of the processing executed by the CPU 101 is performed by a processor other than the CPU 101, such as a dedicated processor such as a digital signal processor (DSP), graphics processing unit (GPU), numerical processor, vector processor, or image processor. may Also, at least part of the processing executed by the CPU 101 may be executed by an integrated circuit (IC) or other digital circuits. Also, at least part of the CPU 101 may include an analog circuit. Integrated circuits include Large Scale Integrated Circuit (LSI), Application Specific Integrated Circuit (AS
ICs), including programmable logic devices (PLDs). PLDs include, for example, Field-Programmable Gate Arrays (FPGAs). CPU 101 may be a combination of a processor and an integrated circuit. A combination is called, for example, a microcontroller unit (MCU), a system-on-a-chip (SoC), a system LSI, a chipset, or the like.

In the information processing apparatus 100, the CPU 101 expands the program stored in the auxiliary storage unit 103 into the work area of the main storage unit 102, and controls the peripheral devices through the execution of the program. Accordingly, the information processing apparatus 100 can execute processing that meets a predetermined purpose. The main storage unit 102 and the auxiliary storage unit 103 are recording media readable by the information processing apparatus 100 . The main storage unit 102 is exemplified as a storage unit directly accessed from the CPU 101 . The main memory unit 102 includes Random Access Memory (RAM) and Read Only Memory (ROM).

The auxiliary storage unit 103 stores various programs and various data in a recording medium in a readable and writable manner. The auxiliary storage unit 103 is also called an external storage device. The auxiliary storage unit 103 stores an operating system (OS), various programs, various tables, and the like. The OS includes a communication interface program for exchanging data with an external device or the like connected via the communication unit 104 . External devices and the like include, for example, other information processing devices and external storage devices connected via a computer network or the like. Note that the auxiliary storage unit 103 may be part of a cloud system, which is a group of computers on a network, for example.

The auxiliary storage unit 103 is, for example, an erasable programmable ROM (EPROM), a solid state drive (Solid State Drive, SSD), a hard disk drive (Hard Disk
Drive, HDD) and the like. Further, the auxiliary storage unit 103 is, for example, a Compact Disc (CD) drive device, a Digital Versatile Disc (DVD) drive device, a Blu-ray (registered trademark) Disc (BD) drive device, or the like. Also, the auxiliary storage unit 103 may be provided by Network Attached Storage (NAS) or Storage Area Network (SAN).

The communication unit 104 is, for example, an interface with a computer network such as the Internet or Local Area Network (LAN). A communication unit 104 communicates with an external device or the like via a computer network.

The input/output interface 105 is an interface with an input/output device, such as a PS/2 connector, a Universal Serial Bus (USB) connector, a Video Graphics Array (VGA) connector, a Digital Visual Interface (DVI) connector, a High-Definition Multimedia Interface (HDMI (registered trademark)) and the like.

A display 210 is an output unit that outputs data processed by the CPU 101 and data stored in the main storage unit 102 . The display 210 is, for example, a Cathode Ray Tube (CRT) display, a Liquid Crystal Display (LCD), a Plasma Display Panel (PDP), an Electroluminescence (EL) panel, an organic EL panel, or the like. The display 210 is connected to the information processing device 100 via the input/output interface 105 .

The keyboard 220 and the mouse 230 are input means for receiving operation instructions and the like from the user. Keyboard 220 and mouse 230 are connected to information processing apparatus 100 via input/output interface 105 .

<Functional Blocks of Information Processing Apparatus 100>
FIG. 2 is a diagram showing an example of functional blocks of the information processing apparatus 100 according to the first embodiment. The information processing apparatus 100 includes a synonym expansion unit 301, a synonym dictionary 301a, a text search unit 302, a text database (denoted as text DB in the figure) 302a, a morphological analysis unit 303, a document vector generation unit 304, and a stop word dictionary. 304a, word vector generation unit 305, word classification degree calculation unit 307, classifier 308, feature model 308a, classification degree vector generation unit 309, base point determination unit 310, display data generation unit 311, proximity rule list 311a, specific feature data list 311b, an attribute data list 311c, a display target data list 311d, a word feature amount comparison unit 312, a node document search unit 313, an edge document search unit 314, a text display unit 315, and a detailed data search unit 316. The information processing apparatus 100 executes the processes of the respective units described above by the CPU 101 executing a computer program that has been developed in the main storage unit 102 so as to be executable.

A large number of documents are stored in the text database 302a. A document is data including at least text, and examples thereof include articles, technical data, specifications, patent documents, books, laws and regulations, contracts, judicial precedents, and documents written in HTML or XML. Documents may include images, moving images, and audio in addition to text. In this specification, the term "document" is used to mean document data or document file, but depending on the context, the term "document" may be used to mean text included in document data or document file. . The text database 302a manages documents in association with document IDs. A document ID is identification information for uniquely identifying a document. Note that if the document is a resource that exists on a network such as the Internet, a Uniform Resource
The Identifier (URI) may be stored in the text database 302a. The text database 302a is an example of a "database."

The text search unit 302 extracts a plurality of documents matching the search conditions from the text database 302a based on the search conditions specified via input means such as the keyboard 220. FIG. A document extracted by the text search unit 302 is called an "extracted document". A search condition includes at least a keyword and may further include a logical operator. The text search unit 302 may receive synonyms of the keywords included in the specified search condition from the synonym expansion unit 301 and execute an OR search specifying the keyword and synonyms. By executing an OR search specifying a keyword and its synonyms, the text search unit 302 can extract documents containing at least one of the keyword specified by the user and its synonyms from the text database 302a. can. The text search unit 302 stores the document ID of the extracted document in the main storage unit 102 and auxiliary storage unit 103 . The text search unit 302 is an example of a "search unit".

The synonym expansion unit 301 searches for synonyms of the designated keyword. Here, synonyms are not limited to words having the same meaning as the keyword, and may be words having similar meaning to the keyword. The synonym developing unit 301 searches the synonym dictionary 301a based on the designated keyword. FIG. 3 is a diagram showing an example of the configuration of the synonym dictionary 301a. Each row of the synonym dictionary 301a indicates a synonym group composed of a plurality of synonyms, and lists one representative word selected from the synonym group and the remaining synonyms. The synonym development unit 301 searches the designated keyword from the synonym dictionary 301a, and returns words included in the synonym group to which the keyword belongs (excluding the same words as the keyword) as synonyms. For example, when "apple" is specified as a keyword, the synonym expansion unit 301 searches for "apple" from the synonym dictionary 301a and finds "apple", "apple", and "apple" included in the synonym group to which "apple" belongs. Returns "Apple" as a synonym for "apple".

The morphological analysis unit 303 performs morphological analysis to divide the text included in the input document into words. The morphological analysis unit 303 divides the text into words based on, for example, a dictionary that associates words with parts of speech, and derives part of speech information corresponding to the words. FIG. 4 is a diagram showing an example of a morphological analysis result by the morphological analysis unit 303. As shown in FIG. FIG. 4 is an example of the result of executing morphological analysis on the text "Apple is a fruit grown in Aomori." In FIG. 4, the left end of each line indicates the divided word. On the right side of the divided word, part of speech, base form, type of conjugation, phonetic notation, etc. are shown as part of speech information of the word, separated by commas.

The morphological analysis unit 303 analyzes the text included in each of the multiple extracted documents received from the text search unit 302 to extract words that appear at least once in the multiple extracted documents. The morphological analysis unit 303 assigns a word ID to each of a plurality of words extracted from a plurality of extracted documents, and stores them in the main storage unit 102 as analysis results. A word ID is identification information for uniquely identifying a word. The morphological analysis unit 303 is an example of an “analysis unit”. In the present embodiment, morphological analysis was exemplified as a specific example of the analysis unit, but the document analysis method is not limited to morphological analysis, and other methods may be employed. For example, in the case of a Japanese document, syntactic analysis including chunking processing may be used in addition to morphological analysis. It is also preferable to use a tokenizer or chunker for English documents.

The morphological analysis unit 303 may extract all words contained in the extracted document, but in order to reduce the number of words to be extracted, the morphological analysis unit 303 may extract only certain parts of speech (for example, nouns), It is also possible to extract only words that are more than a threshold value, or to extract a predetermined number of words from words that appear more often. The morphological analysis unit 303 may also use syntactic analysis to select compound words forming a word or phrase to be extracted, or pairs of compound words forming words or phrases having a dependency relationship. For example, by using syntactic analysis including chunking processing, it is possible to extract semantically coherent compound words and phrases. Also, the morphological analysis unit 303 may generate word N-grams from the result of the morphological analysis. In this case, the character string finally output by the morphological analysis unit 303 is not a "word" but a "compound word or phrase consisting of a plurality of words". There is no particular need to distinguish between and "phrases" or to change processing depending on whether it is a "word", a "compound word" or a "phrase". Therefore, in the following description, the expression "word" is used for convenience. ” or “compound word”. It is preferable to allow the user to specify settings such as the threshold for the number of appearances, the number of words to be extracted, the parameter N in the word N-gram, and the like. Note that when generating word N-grams, only phrases composed of N words may be extracted, or phrases composed of N words or less may be extracted.

A document vector generation unit 304 generates a document vector for each of the multiple words extracted by the morphological analysis unit 303 . A document vector is a vector whose element is the number of appearances of the word in each extracted document. The document vector generation unit 304 stores the generated document vector in the main storage unit 102 or the auxiliary storage unit 103 in association with the word ID. FIG. 5 is a diagram showing an example of the document vector 3041. As shown in FIG. Each column in FIG. 5 indicates a document vector 3041, and each row indicates an extracted document. The numbers in the table indicate the number of times the word in the corresponding column appears in the document in the corresponding row. If the number of extracted documents is M, the document vector 3041 will be an M-dimensional vector. For example, in FIG. 5, the document vector 3041 of the word "apple" with the word ID "101" is indicated by {..., 1, 2, 3, 0, 0,...}. According to this document vector 3041, the word "apple" appears once in the document with the document ID "11", twice in the document with the document ID "12", and three times in the document with the document ID "13". It can be seen that it does not appear in the documents of "14" and "15".

A word vector generation unit 305 generates a word vector for each of the multiple extracted documents extracted by the text search unit 302 . A word vector is a vector whose elements are the number of occurrences of each word in the document. The word vector generation unit 305 stores the generated word vector in the main memory unit 102 or the auxiliary memory unit 103 in association with the document ID. FIG. 6 is a diagram showing an example of the word vector 3051. As shown in FIG. Each row in FIG. 6 indicates a word vector 3051, and each column indicates a word. The numbers in the table indicate the number of times the word in the corresponding column appears in the document in the corresponding row. If the number of words is L, the word vector 3051 becomes an L-dimensional vector. For example, in FIG. 6, the word vector 3051 of the document with the document ID "12" is indicated by {..., 2, 1, 0, 0, 0, 0, 0,...}. With this word vector 3051, the word "apple" appears twice and the word "orange" once in the document with the document ID "12", and the words "tomato", "watermelon", "melon", "cucumber", and "strawberry" appear. does not appear.

The stop word dictionary 304a is a dictionary that stores information indicating words that are not included in the document vector 3041 or the word vector 3051. FIG. FIG. 7 is a diagram showing an example of a stopword dictionary. The document vector generation unit 304 and the word vector generation unit 305 may generate the document vector 3041 and the word vector 3051 by excluding words stored in the stop word dictionary 304a, for example.

A classifier 308 classifies an input document into n classes (n is an integer equal to or greater than 2). The classifier 308 uses, for example, a feature model 308a prepared in advance to calculate and output the score of the input document. This score is a value representing the probability or likelihood that an input document belongs to a certain class, and takes continuous values (the classifier 308 may therefore be called a regressor). For example, in the case of a two-class classifier that classifies input documents into whether they are “documents about fruit” or not, it is preferable to design or learn to output scores in the range of 0 to 1. In this case, the closer the output score is to 1, the more likely it is that the input document is a document about fruit. I can judge. In the case of a three-class classifier that classifies input documents into "documents about vegetables,""documents about fruits," and "other documents," the range is -1 (vegetables) to 0 to +1 (fruits). should be designed or learned to output a score of In this case, the closer the output score is to -1, the more likely it is that the input document is about vegetables, and the closer the output score is to +1, the more likely it is that the input document is about fruits. If the output score is close to 0, it can be determined that "there is a high possibility that the input document is neither a document about vegetables nor a document about fruits." Such a classifier 308 may be created by machine learning using a large amount of teacher data (document samples for training), or may be created based on rules and models designed by humans. Any machine learning method can be used, such as support vector machine (SVM), Bayesian network, neural network (NN), deep neural network (DNN), and the like. SVM is used in this embodiment. Since the output score of the classifier 308 can be said to be a numerical representation of the features of the input document, it is hereinafter referred to as a "document feature score". The classifier 308 is an example of a “document feature calculator” that calculates a document feature score for each extracted document.

Both the word classification degree calculation unit 307 and the classification degree vector generation unit 309 are functions for calculating a feature amount representing the feature of the word based on the document vector 3041 of the word and the document feature score of each document. The difference between the word classification degree calculation unit 307 and the classification degree vector generation unit 309 is that the feature amount (classification degree) obtained by the former is an index (scalar) consisting of one value, whereas the feature amount ( The classification vector) is an index (vector) consisting of a set of multiple values. Each feature amount is an index representing the feature of a word (character string), and is an example of a "character string feature amount." A specific calculation method for each feature amount will be described below.

The word classification degree calculation unit 307 identifies extracted documents in which the word appears one or more times (hereinafter referred to as “appearing documents”) from the document vector 3041 of the target word, and calculates the document features of each of the identified appearing documents. A feature amount of the word is calculated based on the score. Specifically, the word classification degree calculation unit 307 calculates the product of the document feature score of the appearing document and the number of appearances of the word in the appearing document, and calculates the product of the document feature score and the appearance number for all the appearing documents. The total value is used as the feature amount of the word. Since this feature amount is used for word classification in the subsequent tree structure generation processing, this feature amount is referred to as a "word classification degree" in this specification. For example, in the case of "watermelon" in FIG. are "6" and "3". Therefore, the classification of "watermelon" is
Classification degree of "watermelon" = 6 x 0.8 + 3 x (-0.1) = 4.5
Asked. Note that in the present embodiment, the total value of the product of the document feature score and the number of occurrences is defined as the degree of classification, but another statistic may be used instead of the total value. For example, the classification degree may be determined by means, standard deviation, or the like.

The classification vector generation unit 309 identifies the appearing document from the document vector 3041 of the target word, and calculates the feature amount of the word based on the document feature score of each of the identified appearing documents. Specifically, the classification vector generation unit 309 sets a vector having as an element the product of the document feature score and the number of appearances of the word, as the feature amount of the word. Since this feature amount is also used for word classification in the subsequent tree structure generation processing, this feature amount is referred to as a "classification degree vector" in this specification. For example, in the case of "watermelon" in FIG. 8, the classification vector 3091 is {..., 0, 0, 6 x 0.8, 0, 3 x (-0.1),...}. Note that in the example of this embodiment, the degree of classification of a word is equal to the sum of all the elements of the degree of classification vector of the word.

The base point determination unit 310 determines a word that serves as the base point of the tree structure. The word that serves as the base point may be, for example, a word specified by the user, a word with the highest or lowest classification degree, or a word with the highest or lowest classification degree vector 3091 . It can be a small word. Alternatively, the base point determination unit 310 may calculate an average classification degree, which is the average of classification degrees among all words, and select a word having a classification degree closest to the average classification degree among all words as a base point. . In addition, the base point determination unit 310 calculates an average classification vector, which is the average of the classification vectors of all words, and selects the word having the classification vector closest to the average classification vector among all words as the base point. You can choose to The base point determination unit 310 passes information on the word determined as the base point to the display data generation unit 311 . In this embodiment, the magnitude of the classification vector 3091 is defined as "the sum of all the elements of the classification vector". Therefore, in this embodiment, the "word classification degree" and the "word classification degree vector magnitude" have the same value.

Note that the base point of the tree structure may be an empty node. If the base point is an empty node, the base point determination unit 310 selects a pair of the word with the highest classification degree and the word with the lowest classification degree, or the word with the largest classification degree vector and the word with the smallest classification degree vector. A word pair is selected and passed to the display data generation unit 311 .

The display data generation unit 311 generates a tree structure representing the relationship between multiple words and outputs it to the display 210 . In the tree structure generated in this embodiment, each node is associated with a word, and the arrangement of each node is determined based on the difference in feature amount (classification degree or classification degree vector) between words. It is characterized by Details will be described later.

The word feature quantity comparison unit 312 is a function that evaluates the similarity between two words by comparing the feature quantity between the two words. Specifically, the word feature quantity comparison unit 312 calculates the difference in feature quantity between two words and outputs the value as the degree of similarity (in this case, the smaller the difference, the higher the similarity; The higher the value, the lower the similarity). The difference in feature amount can be obtained, for example, as follows. If the feature quantity is the degree of classification (scalar), the difference (subtraction value) in the degree of classification between two words or its absolute value may be calculated. If the feature quantity is a classification vector, the difference in classification vector between two words may be calculated using a vector comparison function such as cosine similarity or Euclidean distance.

When the user designates any node of the tree structure using the keyboard 220, the mouse 230, or the like, the node document search unit 313 retrieves the document ID that is an element of the document vector corresponding to the word of the node. Search for documents. Further, when the user designates any edge of the tree structure using the keyboard 220, the mouse 230, or the like, the edge document search unit 314 retrieves the document vector between two nodes connected by the designated edge. A document with a document ID that is an element of a vector obtained by ANDing is retrieved. An "edge" is a link connecting between nodes. The text search results from the node document search unit 313 and the edge document search unit 314 are passed to the text display unit 315 . A text display unit 315 has a function of displaying the text search result on the display 210 .

When the user designates any node in the tree structure using the keyboard 220, the mouse 230, or the like, the detailed data search unit 316 searches for the detailed data of the word associated with the designated node. It is a function to display. The detailed data search unit 316 searches, for example, a database or a website built in an environment accessible to the information processing apparatus 100 using the word displayed at the designated node as a search key, and displays the search results. The database or website built in the environment accessible by the information processing device 100 may be a database or website built on the Internet, or a database or website built on the auxiliary storage unit 103 of the information processing device 100. It may be a database, a database built on a LAN, or a website.

<Processing flow>
A processing flow executed by the information processing apparatus 100 according to the first embodiment will be described with reference to FIGS. 17 to 20. FIG. 17 to 20 are diagrams showing an example of the processing flow according to the first embodiment. "A" in FIG. 17 connects to "A" in FIG. 18, "B" in FIG. 18 connects to "B" in FIG. 19, and "C" in FIG. 19 connects to "C" in FIG. .

In step S1, search conditions are designated by an input means such as the keyboard 220, and a search query is generated. The search query is passed to text searcher 302 . In step S<b>2 , the text search unit 302 passes the keyword included in the search query to the synonym expansion unit 301 . The synonym development unit 301 acquires synonyms of the received keyword from the synonym dictionary 301 a and passes them to the text search unit 302 .

In step S3, the text search unit 302 extracts documents containing the keyword or synonyms included in the search query from the text database 302a. The processing from step S1 to step S3 is an example of a "search step."

In step S<b>4 , the morphological analysis unit 303 extracts a plurality of words (character strings) by performing morphological analysis on each text of the extracted document obtained by the text search unit 302 . Step S4 is an example of an "analysis step."

In step S5, the document vector generator 304 generates a document vector 3041 for each word obtained by the morphological analyzer 303. FIG. In step S6, the word vector generation unit 305 generates word vectors 3051 for each extracted document obtained by the text search unit 302. FIG. The order of step S5 and step S6 may be exchanged.

At step S7, the classifier 308 calculates a document feature score for each extracted document obtained by the text search unit 302. FIG. In step S8, the word classification degree calculator 307 calculates the classification degree of each word. In step S9, the classification vector generation unit 309 calculates the classification vector of each word. Step S7 is an example of a "document feature calculation step", and steps S8 to S9 are examples of a "character string feature amount calculation step".

In step S10, the base point determination unit 310 determines a word that becomes a base node of the tree structure. The base point determination unit 310 passes the word determined as the base node to the display data generation unit 311 . Note that when the base node is set to "empty", the base point determination unit 310 creates a pair of a word with the highest classification degree and a word with the lowest classification degree vector, or a word with the largest classification degree vector and a word with the smallest classification degree vector. pair to the display data generator 311 . Step S10 is an example of a "base point determination step".

In step S11, the display data generation unit 311 sets the word passed from the base point determination unit 310 as a base node. When the base node is "empty", the display data generation unit 311 arranges the word pairs received from the base point determination unit 310 under the "empty" base node. FIG. 9 is a diagram showing an example of a state in which a node of the word "apple" with the highest classification degree and a node of the word "tomato" with the lowest classification degree are arranged under the base node of "empty". A base point of the tree structure is generated in step S11.

In step S12, the display data generation unit 311 selects words that are candidates to be added to the tree structure next from the remaining words (that is, words that have not yet been arranged in the tree structure). If the base node is "empty", for example, a pair of the word with the highest word classification degree and the word with the lowest word classification degree, or a word with the largest word classification degree vector and the word with the lowest Choose small word pairs. If the base node is not "empty", for example, a word that is most similar to the word of the base node should be selected from among the remaining words (the degree of similarity between words is the same as that of the word feature quantity comparison unit 312). method to calculate). The selected additional candidate words are passed to the word feature quantity comparison unit 312 .

In step S13, the word feature quantity comparison unit 312 identifies nodes to which child nodes can be added, among the nodes already displayed in the tree structure. Since this embodiment deals with a binary tree, a node to which a child node can be added is a node that has no child node or only one child node. Then, the word feature amount comparison unit 312 compares the feature amounts between words for all combinations of the addition candidate word selected in step S12 and the word associated with the node to which the child node can be added, A combination with the highest degree of similarity between words (the smallest difference in feature amounts) is selected. Information on the additional candidate words and nodes to which child nodes can be added is passed to the display data generation unit 311 .

In step S14, the display data generation unit 311 adds a new child node to a node to which a child node can be added, and associates the addition candidate word with the child node. As a result, words with similar feature amounts are linked as child nodes. FIG. 10 is an example of a state in which similar nodes are added. In FIG. 10, a child node "orange" is added under the node "apple", and a child node "cucumber" is added under the node "tomato". Since a binary tree is displayed in this embodiment, a child node is not added to a node having two child nodes.

In step S15, the display data generation unit 311 checks whether there are any unprocessed words (that is, words that have not been added to the tree structure). If unprocessed words remain, the processes of steps S12 to S14 are repeated. If there is no unprocessed word, the process moves to step S16. In step S16, the display data generator 311 outputs the binary tree with the determined structure to a display device such as the display 210 or the like.

The processing after step S17 is processing in response to an operation on the displayed tree structure. At step S17, it is determined whether or not a node has been designated by input means such as the keyboard 220 or the mouse 230 or not. If a node has been designated (YES in step S17), the process proceeds to step S18. If no node is specified (NO in step S17), the process proceeds to step S19.

In step S18, the node document search unit 313 searches the text of the document using the word associated with the specified node as a search key. At this time, the node document search unit 313 identifies documents in which the word appears from the document vector or classification degree vector of the word associated with the specified node, and performs a text search using those documents as a search range. . Text display unit 315 displays the text search results on display 210 . The text search results can be displayed in any manner. For example, the text part containing the word used as the search key may be extracted and displayed near the designated node (at this time, the word used as the search key may be highlighted). . Alternatively, a list of documents (a list of document IDs, a list of document titles and file names, etc.) containing words used as search keys may be displayed.

In step S19, it is determined whether or not an edge has been designated by input means such as the keyboard 220 or the mouse 230 or not. If an edge is specified (YES in step S19), the process proceeds to step S20. If no edge is specified (NO in step 19
), the process proceeds to step S21.

In step S20, the edge document search unit 314 identifies documents appearing from the vector obtained by ANDing the document vector or the classification degree vector of two words respectively associated with two nodes connected by the specified edge. Then, text search is performed with those documents as the search range. Text display unit 315 displays the text search results on display 210 . The text search results can be displayed in any manner. For example, the text part containing the word used as the search key may be extracted and displayed near the designated node (at this time, the word used as the search key may be highlighted). . Alternatively, a list of documents (a list of document IDs, a list of document titles and file names, etc.) containing words used as search keys may be displayed.

In step S21, it is determined whether or not an operation for displaying detailed information has been performed using input means such as the keyboard 220 and the mouse 230 or the like. The operation for displaying the detailed information is, for example, pressing a predetermined key on the keyboard 220, right-clicking the mouse 230, or the like. If an operation for displaying detailed information has been performed (YES in step S21), the process proceeds to step S22. If an operation for displaying detailed information has not been performed (NO in step S21), the process proceeds to step S17.

In step S22, the detailed data search unit 316 displays detailed information of the word associated with the specified node. FIG. 16 is a display example of search results by the detailed data search unit 316. In FIG. FIG. 16 illustrates a state in which the "apple" node is specified, the detailed data search unit 316 searches websites on the Internet using the specified "apple" as a search key, and the search results are displayed. ing.

<Processing example of the display data generator>
Next, variations of tree structure display processing by the display data generation unit 311 will be described. The following display processing is not an essential function, and may be implemented in the display data generation unit 311 as necessary.

(attribute data list)
The display data generation unit 311 may display the corresponding nodes of the words registered in the attribute data list 311c with specific colors and effects. FIG. 11 is a diagram showing an example of the attribute data list 311c. In the attribute data list 311c, a word is associated with an attribute to be assigned to a node associated with the word. In FIG. 11, the background color of the node is specified as an example of the attribute. For example, the node for the word "orange" has a background color of "orange".

(Specific feature data list)
FIG. 12 is a diagram showing an example of the specific feature data list 311b. The specific feature data list 311b is a list of words to which attributes are assigned. Here, the words registered in the specific feature data list 311b are called "target words", and the nodes associated with the target words are called "target nodes". A node adjacent to the target node (that is, a parent node or a child node) is called an "adjacent node", and a word associated with the adjacent node is called an "adjacent word".

The display data generation unit 311 displays the target node and/or the adjacent node with specific colors and effects when the target word and the adjacent word satisfy predetermined conditions. The predetermined condition is, for example, that the degree of similarity between the target word and adjacent words is higher than a predetermined threshold. The degree of similarity between words may be obtained from the difference in word classification, as described above, or may be obtained from the cosine similarity of classification degree vectors, the Euclidean distance, or the like. Moreover, it is preferable that the setting of the predetermined threshold can be changed by the user.

(proximity rule list)
The display data generator 311 may change the node display mode based on the rules registered in the proximity rule list 311a. FIG. 13 is a diagram showing an example of the proximity rule list 311a. The proximity rule list 311a includes four items: rule ID, adjacent color, count, and attribute. The rule ID is an ID that uniquely identifies each rule registered in the proximity rule list 311a. Adjacent color is the background color of adjacent nodes. The number is the number of adjacent nodes. The attribute is the background color of the node that changes when the adjacency color and count conditions are met. For example, in the case of the proximity rule with rule ID: 4, the background color of a node that has two adjacent nodes with a background color of "red" is changed to "yellow." FIG. 14 is a diagram illustrating a case where the node display mode is changed based on the rules registered in the proximity rule list 311a. FIG. 14A illustrates the state before change, and FIG. 14B illustrates the state after change. In FIG. 14, two nodes (a parent node and a child node) adjacent to the node 1001 are red, so the color of the node 1001 is changed to yellow according to the proximity rule of rule ID:4.

The display data generator 311 may refer to the display target data list 311d further, and may not output words that are not registered in the display target data list 311d as nodes. FIG. 15 is a diagram showing an example of the display target data list 311d. Words to be displayed as nodes are registered in the display target data list 311d. In FIG. 15, "orange", "watermelon" and "cucumber" are registered in the display target data list 311d, and words other than these three words are not output as nodes. By excluding words not desired to be displayed as nodes from the display target data list 311d, the display data generation unit 311 can display a tree structure with less noise.

<First modification>
In the first embodiment, the document vector generator 304 generated the document vector 3041 for each word and did not consider synonyms. However, the document vector generation unit 304 may generate the document vector 3041 by referring to the synonym dictionary 301a and replacing the synonym with a representative word selected from the synonyms. FIG. 21 is a diagram showing an example of a document vector 3041a in which synonyms are collectively generated. The upper part of FIG. 21 is an example of a document vector 3041 before synonyms are replaced with representative words, and the lower part of FIG. 21 is an example of a document vector 3041a after synonyms are replaced with representative words. In the upper part of FIG. 21, for example, different document vectors 3041 are generated for the words "ringo", "ringo" and "ringo" indicating the same "ringo". Also, different document vectors 3041 are generated for the words "suika" and "watermelon" indicating the same "suika". In the lower part of FIG. 21, as a result of replacing the synonyms with representative words, "apple", "apple" and "apple" are replaced with the representative word "apple" to generate a document vector 3041a. It can be seen that the document vector 3041a is generated by replacing "watermelon" with the representative word "watermelon". The number of appearances after replacing the synonyms with the representative words is the sum of the number of appearances of each synonym. When the document vector 3041a is generated by replacing the synonyms with the representative words, each node displayed in the tree structure may also be displayed by replacing the synonyms with the representative words. By replacing synonyms with representative words, it is possible to absorb variations in spelling of words in each document.

<Second modification>
In the first embodiment, the text search unit 302 extracted all documents containing keywords given as search conditions from the text database 302a. However, if a date and time is specified as a search condition, the text search unit 302 searches for documents whose creation date matches the specified date and time, documents created after the specified date and time, or documents created before the specified date and time. You may extract the created document. In this case, each document is attached with date-of-creation information indicating the date of creation of the document, and the text search unit 302 can compare the date-of-creation with the designated date and time. When date and time are specified as a search condition, the display data generator 311 may change the node display mode based on the specified date and time and creation date information attached to the document, for example. For example, the display data generation unit 311 displays word nodes that appear in documents whose creation date information matches the specified date and time in a manner different from other nodes, so that the user can easily distinguish them from other nodes. can be displayed as

<Third modification>
In the first embodiment, the document vector generation unit 304 generated the document vector 3041 using all words included in the document except for the words registered in the stop word dictionary 304a. However, the number of words used in the document vector 3041 may be limited by setting a threshold. For example, it is possible to adopt a method in which words are adopted as elements of the document vector 3041 in order of appearance in the document, and the remaining words are not adopted as elements of the document vector 3041 when the number of adopted words reaches a threshold. Also, when the number of words adopted as elements of the document vector 3041 reaches a threshold value, the oldest words adopted as elements are removed from the document vector 3041, and new words are adopted as elements of the document vector 3041. A method is also possible. In addition, the words are weighted by a predetermined algorithm (for example, TF-IDF, Okapi BM25, etc.) for weighting words, and the words with the highest weights are adopted as elements of the document vector 3041 in descending order. A method is also possible in which, when the number of words adopted as the document vector 3041 reaches a threshold value, the subsequent words are not adopted as the elements of the document vector 3041 .

<Fourth modification>
A document may contain references to other documents. For example, if the document is a paper, it may contain information indicating the cited paper as reference information to other documents. In such a case, the text search unit 302 may tally the number of times each document is referenced by other documents, and preferentially extract documents with a large number of times of reference. For example, when a threshold is set for the number of documents to be extracted, the extracted documents are extracted in descending order of the number of references from among the documents extracted based on the search query. When the number of reaches a threshold value, the extraction should be terminated. A document with a large number of references is highly likely to be a document of high importance. Therefore, by adopting such a configuration, there is an advantage that documents from which there is a high possibility of obtaining important information are preferentially extracted.

Further, the text search unit 302 may tally the number of times each document is referred to by other documents, and preferentially extract documents with a low number of times of referencing. The priority extraction in this case means, for example, when a threshold is set for the number of documents to be extracted, the documents extracted based on the search query are extracted in descending order of the number of references, and When the number of extracted documents reaches the threshold, the extraction should be terminated. A document with a small number of references is highly likely to be a document containing information that is not well known. Therefore, by adopting such a configuration, there is an advantage that documents that are highly likely to provide new knowledge to the user are preferentially extracted.

<Fifth Modification>
In the first embodiment, the initial base point of the tree structure is determined by the base point determination unit 310, but the user may change the base point after the tree structure is displayed. For example, when the user designates a node to be used as a base point using input means such as the keyboard 220 or the mouse 230, the base point determination unit 310 passes the designated node as a base point to the display data generation unit 311, and the display data generation unit 311 A binary tree can be generated with the specified node as the base point. The base point determination unit 310 and the display data generation unit 311 that execute such processing are an example of the “base point change unit”.

<Sixth modification>
In the first embodiment, a binary tree was exemplified, but the tree structure may be a ternary tree or a tree structure that branches into more branches. In this case, the user may specify the number of branches for the display data generator 311 via input means such as the keyboard 220 . For example, when the tree structure is a ternary tree, "3" may be specified as the number of branches.

<Seventh Modification>
In the first embodiment, when the base node is "empty", the word with the maximum classification degree or the magnitude of the classification degree vector (hereinafter collectively referred to as "classification degree") is used as a node connected under the base point. The minimum word pair is selected (see step S10), and as nodes to be added thereafter, a word pair with the highest and lowest classification degrees is selected from the remaining words (see step S12). ). Such a selection procedure is a suitable example when the tree structure is a binary tree and the degree of classification is an index representing "whether or not the word belongs to a certain class". If the tree structure is a binary tree and the classification degree is an index indicating whether the word belongs to the first class or the second class, then in steps S10 and S12, the first The word with the highest degree of classification to the class of 1 and the word with the highest degree of classification to the second class should be selected. Further, when the tree structure is a ternary tree and the classification degree is an index representing "whether the word belongs to the first class, the second class, or the third class", In steps S10 and S12, three words are selected: the word with the highest degree of classification into the first class, the word with the highest degree of classification into the second class, and the word with the highest degree of classification into the third class. Just do it. The same applies when the number of branches is more than three.

<Advantages of the First Embodiment>
The advantages of the first embodiment described above are summarized as follows. In the tree structure described above, the arrangement of each node is determined based on the difference in the feature amount (classification degree or classification degree vector) representing the feature of the word. It is possible to easily grasp the tendencies of words that appear in multiple extracted documents that are search results. Moreover, in the above embodiment, the feature of a word is expressed using the document feature score, which is the feature of the document (text, context) using the word, rather than the word itself. Therefore, the arrangement and connection relationship of each node in the tree structure reflect the relationship and similarity between documents. Therefore, by using the above-described tree structure, it is possible to appropriately and intuitively express the relationship/similarity between documents and the relationship between words appearing in a plurality of documents. It becomes possible to support

<Second embodiment>
A second embodiment of the present invention will be described with reference to FIG. In the second embodiment, a word classifier (character string classifier) is used to obtain a classification degree, which is a word feature amount, directly from a word.

As shown in FIG. 22, the information processing apparatus 100 according to the second embodiment includes a word classifier 401 and word feature models 401a. Other configurations are the same as those of the first embodiment.

The word classifier 401 is a classifier that classifies input words into n classes (n is an integer equal to or greater than 2). The word classifier 401 uses, for example, a word feature model 401a prepared in advance to calculate and output the score of the input word. This score is a value representing the probability or likelihood that an input word belongs to a certain class, and takes continuous values (therefore, the word classifier 401 may be called a regressor). Such a word classifier 401 may be created by machine learning using a large amount of teacher data, or may be created based on human-designed rules and models. Any machine learning method can be used, such as support vector machine (SVM), Bayesian network, neural network (NN), deep neural network (DNN), and the like. SVM is used in this embodiment.

In the case of machine learning, data of a plurality of documents in which character strings appear may be used as teacher data. By using document data in which the character string appears as teacher data for learning the correspondence relationship between the character string and the character string feature amount, the feature amount (classification degree) obtained by the method of the first embodiment It is possible to obtain a feature quantity with characteristics similar to those of For example, when learning a word classifier that classifies character strings into two categories of "fruit" and "vegetable", a large number of document data describing "fruit" and A large number of written document data are used as teacher data. Then, according to the ratio of occurrences of character strings (e.g., "apple", "orange", etc.) extracted from the training data (that is, documents in the "fruits" category and documents in the "vegetables" category), Then, the likelihood of classifying the character string into each category (that is, “fruit-likeness” and “vegetable-likeness”) is learned. Using such a word classifier, for example, when the character string "apple" is input, an output score such as "fruit: 0.98, vegetable: 0.31" is obtained.

As a method other than the above, a thesaurus such as WordNet may be used to calculate the semantic distance (conceptual distance) between words.

Note that the output score of the word classifier 401 is obtained by quantifying the characteristics of the character string represented by the word, and is an example of the "character string characteristic quantity". Also, the word classifier 401 is an example of a “character string feature calculator”.

<Advantages of Second Embodiment>
With the configuration of the second embodiment described above, it is possible to obtain the same effects as those of the first embodiment.

<Computer-readable recording medium>
An information processing program that causes a computer or other machine or device (hereinafter referred to as a computer or the like) to implement any of the functions described above can be recorded in a computer-readable recording medium. By causing a computer or the like to read and execute the program of this recording medium, the function can be provided.

Here, a computer-readable recording medium is a recording medium that stores information such as data and programs by electrical, magnetic, optical, mechanical, or chemical action and can be read by a computer, etc. Say. Examples of such recording media that can be removed from a computer or the like include flexible discs, magneto-optical discs, Compact Disc Read Only Memory (CD-ROM), Compact Disc-Recordable (CD-R), Compact Disc-ReWriterable. (CD-RW), Digital Versatile Disc (DVD), Blu-ray Disc (BD), Digital Audio Tape (DAT), 8 mm tape, and memory cards such as flash memory. In addition, there are a hard disk, a ROM, and the like as recording media fixed to a computer or the like.

DESCRIPTION OF SYMBOLS 100... Information processing apparatus 210... Display 220... Keyboard 230... Mouse 3041... Document vector 3051... Word vector 3091... Classification degree vector

Claims (14)

a search unit for extracting a plurality of documents that match a search condition as extracted documents from a group of documents accumulated in a database;
an analysis unit that analyzes the plurality of extracted documents to extract a plurality of character strings from the plurality of extracted documents as extracted character strings;
a document feature calculation unit for calculating a document feature score obtained by quantifying the feature of each of the plurality of extracted documents;
a character string feature calculation unit for obtaining, for each of the plurality of extracted character strings, a character string feature quantity representing the feature of the extracted character string from document feature scores of one or more extracted documents containing the extracted character string;
an output unit that outputs a tree structure in which each of the plurality of extracted character strings is associated with a node and each node is arranged based on the difference in character string feature amounts between the extracted character strings;
has
The character string feature calculation unit uses a vector having, as elements, values based on document feature scores of one or more extracted documents containing the extracted character string, as the character string feature amount of the character string.
Information processing equipment. The value based on the document feature score is a value obtained by multiplying the document feature score of the extracted document by the appearance frequency of the extracted character string in the extracted document,
The information processing device according to claim 1 . The document feature calculation unit includes a classifier that classifies an input document into n classes (n is an integer equal to or greater than 2). be the document feature score,
The information processing apparatus according to claim 1 or 2 . a search unit for extracting a plurality of documents that match a search condition as extracted documents from a group of documents accumulated in a database;
an analysis unit that analyzes the plurality of extracted documents to extract a plurality of character strings from the plurality of extracted documents as extracted character strings;
A character string feature calculation unit for obtaining, for each of the plurality of extracted character strings, a character string feature amount representing a feature of the extracted character string, wherein the input character string is divided into n classes (n is an integer equal to or greater than 2). A character string feature calculation unit configured by a character string classifier for classification, and using an output score when the extracted character string is input to the character string classifier as a character string feature amount of the extracted character string;
an output unit that outputs a tree structure in which each of the plurality of extracted character strings is associated with a node and each node is arranged based on the difference in character string feature amounts between the extracted character strings;
Information processing device having The search condition includes a keyword,
The search unit acquires synonyms of the keyword from a synonym dictionary that stores synonyms having the same meaning, and extracts documents containing at least one of the keyword and the synonyms as the extracted documents.
The information processing apparatus according to any one of claims 1 to 4 . calculating an average of character string feature amounts among the plurality of extracted character strings, and determining an extracted character string having a character string feature amount closest to the average among the plurality of extracted character strings as a base point of the tree structure; having a base point determination unit;
The information processing apparatus according to any one of claims 1 to 5 . further comprising a base point changing unit that changes the base point of the tree structure to an extracted character string specified by a user;
The information processing apparatus according to any one of claims 1 to 6 . The base point changing unit causes the user to specify any node of the output tree structure, and changes the base point of the current tree structure to the extracted character string associated with the node specified by the user. ,
The information processing apparatus according to claim 7 . The output unit outputs a node corresponding to a predetermined extracted character string, and/or an extracted character string and the predetermined extracted character associated with adjacent nodes that are nodes corresponding to the predetermined extracted character string. Outputting nodes whose columns and predetermined conditions and/or adjacent nodes and nodes satisfying predetermined conditions in a manner different from other nodes,
The information processing apparatus according to any one of claims 1 to 8 . each of the document groups stored in the database includes reference information indicating other documents to be referred to;
The search unit preferentially extracts a document with a small number of references from other documents based on the reference information.
The information processing apparatus according to any one of claims 1 to 9 . each of the document groups stored in the database includes reference information indicating other documents to be referred to;
The search unit preferentially extracts a document with a large number of references from other documents based on the reference information.
The information processing apparatus according to any one of claims 1 to 9 . The analysis unit extracts consecutive words with a predetermined number of words or less as one extracted character string,
The information processing apparatus according to any one of claims 1 to 11 . When any node of the tree structure is specified by the user, the output unit outputs extracted document information including an extracted character string associated with the specified node.
The information processing apparatus according to any one of claims 1 to 12 . When any edge of the tree structure is specified by the user, the output unit outputs an extracted document containing both two extracted character strings respectively associated with two nodes connected by the specified edge. output information,
The information processing apparatus according to any one of claims 1 to 13 .

Images