JPH07120357B2 - Document retrieval device - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document search device for searching a document composed of texts, images and the like.

2. Description of the Related Art A block diagram of a conventional document retrieval apparatus is shown in FIG. This apparatus includes a search condition storage unit 1, a document storage unit 2, a document search unit 3, a document output unit 4, and a search instruction unit 5 for activating the document search unit 3, and when a search is instructed, the search condition storage unit 1 The document corresponding to the key can be searched from the documents in the document storage unit 2. At that time, there is a wildcard search (a prefix match search, a suffix match search, an intermediate match search, etc.) as one of the designations of the search method. The wildcard search is a search method in which a part of the search key in the search condition is designated to search for a document having a key including the search key as a partial character string.

DISCLOSURE OF THE INVENTION Problems to be Solved by the Invention In a conventional document search apparatus, particularly in a search by a wildcard search method, a partial character string is designated as a search key, and a document having a character string including it as a key is searched. It was a thing. Therefore, since the meaning is different from the meaning of the designated character string but the notation is the same, an undesired document has also been searched. For example, when a wildcard search (intermediate match search) is performed by specifying "sex" as the search key, documents that have "difference in physical fitness by sex" and "classification by proper noun attribute" as keys Will be searched for. However, for the purpose of retrieval, the former document should be retrieved, but the latter document should not be retrieved. As described above, since an undesired document may be retrieved, it is necessary to check the retrieved document by visual inspection or the like. Further, there is a problem that the number of retrieved documents becomes unnecessarily large.

In view of the above-mentioned problems of the conventional document search device, the present invention realizes a high-accuracy document search by specifying a search when a partial character string and a wildcard search are specified as a search key. An object of the present invention is to provide a document search device that searches for a document having a key that matches not only the key and the notation but also the morpheme.

Means for Solving Problems The present invention, in order to achieve the above object, a search condition storage unit that stores a search condition that is a condition for designating a document to be searched and that includes at least a search key, A document storage unit that stores a large number of documents and at least a key associated with the document as a pair, a document search unit that retrieves a document from the document storage unit according to the content of the search condition storage unit, a search key and a document A document retrieval device is provided with a morpheme analysis unit for performing a morpheme analysis of the key to be performed and a morpheme match determination unit for comparing the analysis results by the morpheme analysis unit.

With the above-described configuration, the present invention performs morphological analysis of a partial character string designated as a search key in the search condition and a character string of a key attached to the document when searching for a corresponding document based on the search condition. It is compared and judged, and a document having a key whose morphological analysis result matches is searched for.

Example Hereinafter, an example of the present invention will be described with reference to the drawings. First
The figure shows a block diagram of the document retrieval apparatus in the first embodiment of the present invention. In FIG. 1, reference numeral 11 denotes a search condition storage unit for storing information serving as a clue for specifying a document to be searched from a document storage unit described later. In this embodiment, the search condition is search method specification information that is information specifying a plurality of keywords and a logical relationship between keywords such as logical sum and logical product and a search method such as wildcard search. Reference numeral 12 denotes a document search unit, which searches the document storage unit 13 described later for a document having a keyword stored in the search condition storage unit 11 and specifies the document if a corresponding document exists. Is output.
As the designation information, the start address of the area in the document storage unit 13 in which the document is stored is used. here,
When there are a plurality of documents for the search condition, the document search unit 12 outputs a plurality of document designation information. If there is no corresponding document, the document search unit 12 outputs information to that effect. A document storage unit 13 stores a large number of keywords and documents in association with each other. For this, for example, a magnetic disk device, an optical disk device, a magneto-optical disk device, or the like is used. A morpheme analysis unit 14 performs morpheme analysis on the search key in the search condition storage unit 11 and the key in the document storage unit 13 and outputs the analysis result. Reference numeral 15 denotes a morpheme match determination unit that compares the outputs of the morpheme analysis unit 14 and determines whether they match. Reference numeral 17 denotes a search document output unit, which displays a document designated by the document designation information output from the document search unit 12 and which is determined by the morpheme match determination unit 15 to match the search key and the word segmentation of the key. . For this, for example, a display device or a video device is used. If there is no document satisfying the above condition, a message indicating that the document corresponding to the search condition cannot be retrieved in the document storage unit 13 is displayed. A search instruction unit 16 controls the above operation.

The operation of the document searching apparatus of this embodiment configured as described above will be described below.

Search Operation (1) The search instruction unit 16 sets search conditions in the search condition storage unit 11. Here, it is assumed that a wildcard search as a search method, a partial character string of a key as a search key, and a logical relationship between search keys such as logical sum and logical product are set.

(2) The search instruction unit 16 instructs the document search unit 12 to search for a document.

(3) The document search unit 12 searches the document storage unit 13 for a document having a key with each search key stored in the search condition storage unit 11 as a partial character string. For example, as the information for specifying the document, the start address AD of the area in the document storage unit 13 in which the document is stored
Output R. Here, a plurality of the above-mentioned document designation information ADR may be output in response to the search condition. Also,
If there is no corresponding document, the document retrieval unit 12 returns information NO to that effect.
Is output.

(4) The morphological analysis unit 14 morphologically analyzes the search key in the search condition storage unit 11, divides it into word units, and stores it in the internal storage A (not shown). Further, each key associated with the document designated by the output of the document retrieval unit 12 is also subjected to morphological analysis, divided into word units, and stored in another internal storage B (not shown).

(5) The morpheme match determination unit 15 determines whether the content of the internal memory A and the content of the internal memory B of the morpheme analysis unit 14 match. For example, the content of the internal memory A is [... / gender /
...], the content of the internal memory B is [... / no / gender /
In case of / ...], it is determined that they match [... / no /
Attribute / different / classification / ...], it is determined that they do not match.

(6) The search document output unit 17 performs the following processing according to the content of the document designation information for the document having the key determined to match by the morpheme match determination unit 15.

-When the content of information is ADR: The document designated by this is taken out from the document storage unit 13 and temporarily stored, or displayed after being converted into another state such as visual information.

-When the content of information is NO: A message indicating that the document corresponding to the search condition cannot be searched in the document storage unit 13 is displayed.

(End of Search Operation) As described above, according to this embodiment, not only a search key specified as a search condition is included as a partial character string, but a document having only a key whose morpheme matches is additionally searched. be able to. Therefore, it is possible to exclude a search for a document having a key that is semantically different from the specified search condition. Further, the content to be stored in the document storage unit may be the same as that of the conventional document retrieval apparatus and is consistent with the conventional apparatus.

Hereinafter, a second embodiment of the present invention will be described with reference to the drawings.
FIG. 2 is a block diagram of a document search device showing a second embodiment of the present invention. In the figure, 11 is a search condition storage unit, 14 is a morpheme analysis unit, 15 is a morpheme match determination unit, and the above is the same as the configuration of FIG. Reference numeral 12 denotes a document search unit, which searches the document storage unit 13 described later for a document having a keyword stored in the search condition storage unit 11 and specifies the document if a corresponding document exists. Is output. As the designation information, the start address of the area in the document storage unit 13 in which the document is stored is used. Here, if there are a plurality of documents for the search condition, the document search unit 12 outputs a plurality of document designation information. If there is no corresponding document, the document search unit 12 outputs information to that effect. A document storage unit 13 stores a large number of keywords, their morpheme information, and documents in association with each other. For this, for example, a magnetic disk device, an optical disk device, a magneto-optical disk device, or the like is used. A search document output unit 17 displays the document designated by the document designation information output from the document retrieval unit 12. For this, for example, a display device or a video device is used. If there is no document satisfying the above condition, a message indicating that the document corresponding to the search condition cannot be retrieved in the document storage unit 13 is displayed. A search instruction unit 16 controls the above operation. Reference numeral 19 denotes a registration key storage unit, which stores a plurality of registration keys to be stored in association with the document. In the present embodiment, the registration key includes a document title, a creator, etc. in addition to a plurality of keywords. Reference numeral 20 denotes a document registration unit, which is a document, a registration key in the registration key storage unit 19, and a morphological analysis unit.
The morphological analysis results that are the outputs of 14 are associated and stored in the document storage unit 13.

The operation of the document retrieval apparatus of the second embodiment configured as described above will be described below.

Registration Operation (1) The operator attempts to register a document using this apparatus, and sets document information and a registration key by some means. The registration key is stored in the registration key storage unit 19.

(2) The operator instructs the apparatus to register the document.

(3) The morpheme analysis unit 14 morphologically analyzes each keyword in the registration key storage unit 19 and outputs the analysis result.

(4) The document registration unit 20 refers to the document storage unit 13, obtains the head address R of the area where the registered document is to be stored, and the document information (not shown) and the registration key in the registration key storage unit 19 are obtained after R. And the morpheme analysis result of each registration key output from the morpheme analysis unit 14 is registered in the document storage unit 13 in association with each other.

(End of registration operation) Search operation (1) The search instruction unit 16 sets search conditions in the search condition storage unit 11. Here, it is assumed that a wildcard search is set as the search method, a partial character string of the key is set as the search key, and a logical relationship between the search keys such as logical sum and logical product is set.

(2) The search instruction unit 16 instructs the document search unit 12 to search for a document.

(3) The morphological analysis unit 14 morphologically analyzes the search key in the search condition storage unit 11, divides it into word units, and stores it in the internal storage A (not shown).

(4) If the document search unit 12 is accompanied by a key having each search key stored in the search condition storage unit 11 as a partial character string, and the morpheme match determination unit 15 described later matches the morphemes. The document to be judged is searched from the document storage unit 13, and if the corresponding document exists, the head address ADR of the area in the document storage unit 13 in which the document is stored is output as information designating the document. Here, a plurality of the above-mentioned document designation information ADR may be output in response to the search condition. If there is no corresponding document, the document search unit 12
Outputs information NO to that effect.

(5) At that time, the morpheme coincidence determination unit 15 uses the morpheme analysis unit 14
It is determined whether or not the contents of the internal storage A of FIG. 3 and the contents of the morphological analysis result of the registration key stored in association with each document in the document storage unit 13 match. For example, if the content of the internal memory A is [... / Gender / ...], and the content of the morphological analysis result of the registration key is [... / No / Gender / By / ...], it is determined that they match. ... / of / attribute / different / classification /
...], it is determined that they do not match.

(6) The search document output unit 17 performs the following processing according to the content of the document designation information output by the document search unit 12.

-When the content of information is ADR: The document designated by this is taken out from the document storage unit 13 and temporarily stored, or displayed after being converted into another state such as visual information.

-When the content of information is NO: A message indicating that the document corresponding to the search condition cannot be searched in the document storage unit 13 is displayed.

(End of Search Operation) As described above, in the present embodiment, not only the search key specified as the search condition is included as a partial character string, but a document having only a key whose morpheme matches is additionally searched for. You can Therefore, it is possible to exclude a search for a document having a key that is semantically different from the specified search condition. Further, in this embodiment, the morphological analysis of the registration key attached to the document is performed when the document is registered and not performed when the document is searched, so that the document can be searched at high speed.

In the first and second embodiments, the document search section outputs a plurality of document designation information when searching a plurality of documents for given search conditions. It is also possible to output only the designation information of a specific document therein. In the first and second embodiments, the designated address of the retrieved document is the start address of the area in the document storage unit in which the document is stored, but the document such as the document identifier is uniquely identified. Other information to be used may be used. In the first and second embodiments, the search condition is a plurality of keywords and a logical relationship between keywords such as a logical sum and a logical product and search method designation information.
Other information may be included.

EFFECTS OF THE INVENTION In the document search device of the present invention, it is possible to search only a document having a key that matches not only the search key specified as a search condition but also the notation and the morpheme. And the search accuracy is improved. In addition, the number of retrieved documents can be reduced because unsuitable documents are not retrieved. Due to these, the practical effect is great.

[Brief description of drawings]

FIG. 1 is a block diagram of a document search device according to an embodiment of the present invention, FIG. 2 is a block diagram of a document search device according to another embodiment of the present invention, and FIG. 3 is a block diagram of a conventional document search device. is there. 11 ... Search condition storage unit, 12 ... Document search unit, 13 ... Document storage unit, 14 ... Morphological analysis unit, 15 ... Morphological match determination unit.

Claims (1)

[Claims] 1. A search condition storage unit for storing a search condition including a search key, which is a condition for designating a document to be searched, and a large number of documents and at least keys associated with the document are stored as a pair. A document storage unit, a document search unit that searches a document from the document storage unit according to the contents of the search condition storage unit, a morpheme analysis unit that morphologically analyzes a search key and a key associated with the document, and a morpheme analysis unit. A document search device comprising a morpheme match determination unit that compares analysis results with each other, and searches for a document having a key whose morpheme matches a search key.