JP2000200285A - Method and device for information retrieval, and recording medium where information retrieval program is recorded - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make extractible more matching documents from a document group while making the intension of a retriever reflected therein by putting different kind of similarities together and calculating one composite similarity, and presenting a document from a retrieval object according to the composite similarity. SOLUTION: When a retrieval request from the retriever is inputted to a retrieval request input part 14, a similarity calculation part A15 analyzes the retrieval request by a method A, calculates the similarity to a document to be retrieved, and outputs a pair of the similarity and an identifier indicating the document in the form of a list. A similarity calculation part B16 analyzes the retrieval request by a method B and outputs a pair of the similarity to the document to be retrieved and an identifier indicating the document in the form of a list. A similarity composition part 18 puts the similarities together while taking pairs of similarities and identifiers out of similarity calculation results outputted from the similarity calculation parts A15 and B16. Further, a similarity order rearrangement part 19 rearranges the similarity results according to the composite similarity and a retrieval result display part 20 displays the retrieval result.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information retrieval technique, and more particularly, to an information retrieval that reflects the intention of a searcher, and is capable of extracting more documents matching the retrieval intention from a group of documents to be retrieved. The present invention relates to an information search method, an information search device, and a recording medium on which an information search program is recorded.

[0002]

2. Description of the Related Art There are two main types of information retrieval methods in the prior art. A first method is to specify a character string as a search request and extract a document including the character string from a search target document group. In the second method, a document and a search request are first characterized by some criteria, and a similarity, which is a degree of similarity between the characterized documents, is calculated. Next, a method in which a target having a high degree of similarity is set as a target having a high degree of matching, and the target is extracted from a document group to be searched in descending order of degree of similarity.

A typical example of the first method is a character string matching method. That is, this is a method of extracting a document including a character string that matches the character string of the search request. As an example of the second method, there is a tf · idf method. This method statistically analyzes the words contained in a document, uses the frequency of occurrence of the words, regards them as the statistical importance of the words, and uses the importance of the matched words to determine the relevance of the document. It is a way to order. In each of the above examples, it is assumed that the document to be searched includes the character string of the search request.

[0004] As an example of the second method, there is a method in which a character string of a search request may not be included. The method is a method of preparing semantic knowledge about words, characterizing words constituting a document by developing the words with semantic knowledge, and calculating a similarity between the characterized characters.

[0005]

As described above, in the method on the premise that a character string is included, a document containing a character string that exactly matches the character string included in the search request is searched.
In addition, attempts have been made to reduce the omission of search by preparing a synonym dictionary (including a thesaurus) and enabling a search if a synonym is included.

This method of searching for a document containing a character string or a synonym thereof is very effective when the character string of the search request clearly indicates the search intention of the searcher. However, according to this method, there is a characteristic that even if the meaning of the character string and the subject of the document including the character string are irrelevant, the character string can be searched. On the other hand, according to the method of characterizing a document using semantic knowledge, a character string is not necessarily included, and a document related to the meaning of the word is searched even if the word is not a word specified in the search request. However, a document including a character string is not always searched, and this method is not suitable for searching a document including a character string.

The present invention has been made in view of the above points, and an information search method, an information search apparatus, and an information search method for extracting more suitable documents from a group of documents to be searched by reflecting a searcher's intention. It is an object to provide a recording medium on which a search program is recorded.

[0008]

To achieve the above object, the present invention is configured as follows. The present invention sets a plurality of documents as a search target, calculates a similarity between each document in the search target and a word string including one or more words, and calculates one or more documents from the search target according to the similarity. A first calculation step of calculating a plurality of types of the similarity by a plurality of methods for each document;
A second calculating step of calculating one combined similarity by combining the plurality of types of similarities; and presenting the one or more documents from the search target according to the combined similarity.

According to the present invention, the similarity is calculated by a plurality of different methods, and the similarities are combined, whereby
The problem in the related art in which only a single similarity is used is solved, and a search result more suitable for a search intention can be obtained. In the above configuration, the present invention can have the following configuration. The plurality of methods in the first calculation step include a method of calculating a similarity by weight of word association using each document and the word association in the word string, and a method of calculating the similarity in each document. There are two methods of calculating the similarity by checking whether a character string that matches the character string of the word in the column exists.

According to the present invention, the result is synthesized by using two methods, a method in which the character string of the search request does not necessarily match and a method in which only the document in which the character string matches is searched. In addition, the advantages of both methods according to the prior art can be adopted, and it is possible to obtain a search result more suitable for the search intention. In the above configuration, the present invention may have the following configuration.

The second calculating step includes a step of weighting each similarity calculated in the first calculating step, and summing the weighted similarities. According to the present invention, since a plurality of similarities are weighted and combined, it is possible to make use of characteristics that each calculation method is good at.

In the above configuration, the present invention may have the following configuration. The weighting is performed by multiplying the similarity by a coefficient specified from the outside. According to the present invention, the weighting can be performed by designating the weighting from the outside, so that each of the plurality of similarity calculation methods can take advantage of its specialty and reflect the search intention. Becomes

In order to achieve the above object, the present invention may be configured as follows. In the following invention,
It has the same operation and effect as the above invention. The present invention includes a plurality of documents as a search target, and has means for calculating a similarity between each document in the search target and a word string including one or more words. An information retrieval apparatus for presenting a plurality of documents, wherein a plurality of types of the similarity are calculated by a plurality of means for each of the documents.
Calculating means, a plurality of similarities are combined to calculate one combined similarity, and the one or more documents are presented from the search target according to the combined similarity. And

In the above configuration, the present invention can also be configured as follows. The plurality of means in the first calculating means are means for calculating the similarity by the weight of word association using the associative relationship between each document and the word in the word string; Means for calculating the degree of similarity by checking whether there is a character string that matches the character string of the word in the column.

In the above configuration, the information processing apparatus may have the following configuration. The second calculating means includes means for weighting each similarity calculated by the first calculating means, and means for summing the weighted similarities. Further, in the above configuration, the present invention may be configured as follows.

The means for performing weighting includes means for externally designating a coefficient representing the weighting, and means for multiplying the coefficient by the similarity. In order to achieve the above object, the present invention can be configured as follows. According to the present invention, a plurality of documents are to be searched, and
A recording medium for recording an information retrieval program for calculating a similarity with a word string composed of one or more words, and causing a computer to execute a procedure for presenting one or more documents from the search target according to the similarity; A first calculation procedure of calculating a plurality of types of the similarity for each of the documents by a plurality of procedures, a second calculation procedure of calculating one combined similarity by combining the plurality of the similarities, And a step of presenting the one or more documents from the search target according to the composite similarity.

In the above configuration, the present invention can be made as follows. The plurality of steps in the first calculation step include a step of calculating a similarity by a weight of word association using an associative relationship between each of the documents and the words in the word string, and a step of calculating the similarity in each of the documents. And calculating the similarity by checking whether there is a character string that matches the character string of the word in the column.

Further, in the above configuration, the present invention can be configured as follows. The second calculation procedure includes a procedure of weighting each similarity calculated by the first calculation procedure, and a procedure of summing the weighted similarities. Further, in the present invention, the procedure of performing the weighting includes a procedure of externally designating a coefficient representing the weight, and a procedure of multiplying the coefficient by the similarity.

[0019]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing a configuration of an apparatus according to one embodiment of the present invention. As shown in the figure, the information search device of this embodiment includes a search target storage unit 11, an index unit A12, an index unit B13,
Search request input unit 14, similarity calculation unit A15, similarity calculation unit B16, similarity synthesis ratio input unit 17, similarity synthesis unit 1
8, a similarity order rearranging unit 19 and a search result display unit 20 are provided.

Index part A12 / similarity calculation part A1
5 and an index part B13 / similarity calculation part B16 are provided corresponding to the method of calculating the similarity, which is the degree of relevance of the search target document to the search request (hereinafter, referred to as A method and B method). . In this embodiment, two types of methods are used, but any number of types can be used. For example, when n types of methods are used, n sets of index units and similarity calculation units corresponding to each method are provided. In addition, if it is one kind, it is equivalent to an information search device by a single method.

Next, the operation of this embodiment will be described. The search target storage unit 11 stores documents to be searched. The index unit A12 stores an identifier for each search target document and a feature vector used when calculating the similarity in the A method, and the index unit B13 stores an identifier for each search target document and the feature vector used when calculating the similarity in the B method. Is stored.

The search request from the searcher is input to search request input unit 1
4, the similarity calculation unit A15 analyzes the search request by the A method and stores the search request in the index unit A12.
The similarity with the document to be searched is calculated using the index corresponding to the method, and the similarity is output as a list of a pair with the identifier indicating the document. It analyzes and calculates the similarity with the document to be searched using the index corresponding to the method B stored in the index unit B13, and outputs the similarity as a list of pairs with the identifier indicating the document.

The similarity synthesizing unit 18 includes a similarity calculating unit A15
The similarity is combined while extracting the set of the similarity and the identifier from the similarity calculation result output from the similarity calculation unit B16, and the similarity order rearranging unit 19 arranges the similarity results according to the combined similarity. Alternatively, the search result display unit 20 displays the search results. The similarity combination ratio input unit 17 will be described later.

FIG. 2 is a flowchart showing a detailed operation of the similarity synthesizing section 18 shown in FIG. There are various methods for synthesizing similarity, and the method described here is an example of them. In the present embodiment, the method A uses a method of calculating the similarity by the weight of the word association using a knowledge base in which the word association is accumulated, and the method B uses a character string matching method. Used. In this embodiment, the similarity in the character string matching method is set to 1 when a matching character string exists in the search target, and set to 0 when there is no matching character string.

In FIG. 2, first, in step 11, all identifiers and similarities are read from the similarity list output from the similarity calculator A15 in FIG. Next, in step 12, pairs of similarities and identifiers are extracted one by one from the similarity list output from the similarity calculator B16. In step 13, a set having the same identifier as the identifier extracted from the similarity list based on the method B is searched for from the set based on the method A previously read. In step 14, it is checked whether or not the same identifier exists. If not, in step 15, a pair of the identifier and the similarity is added to the end of the previously read list. If there is, in step 16 the similarity between the A and B methods is extracted,
From the similarity of the two methods to calculate the synthetic similarity s ₀ according to Synthesis equation _{F (s 1, s 2)} . The synthesis formula is, for example, s ₀ = α · s ₁ + β · s ₂ . Here, s ₁ is similarity by the A method, and s ₂ is B
The similarity according to the method, α and β, are parameters input from the similarity combination ratio input unit 17 of FIG. By adjusting this parameter, the weight of the characteristic of each method can be changed. The synthesis calculation formula is not limited to the above formula, and may be set to another formula.

In step 17, the similarity of the identifier on the list is replaced with the composite similarity obtained from the above equation. In step 18, it is checked whether or not there is a newly read set in the similarity list by the method B. If there is, the process from step 12 is performed, and if not, the process ends. Next, the present embodiment will be described using a specific example.

FIG. 3 shows an example of an output in the case where the search request is set to “bubble” and the document to be searched is searched by the method A and the similarity is calculated, and is a list including the similarity and the document identifier. From the left of each line, the order, identifier, similarity, and document title are shown. For example, in the first line of FIG. 3, the rank is 1, the identifier is NK950215-2322, the similarity is 0.540805, and the document title is “Japanese President (3) Welfare, bubble is hidden (anything data)”. It is.
The output of the similarity calculation unit A15 in FIG. 1 may include only the identifier and the similarity, but in this embodiment, the order and the document title are also added for explanation.

FIG. 4 shows an output obtained by calculating the similarity by the method B. The structures of the search request and the output result are the same as those in FIG. As described above, since the method B is a character string matching method, only the method including the character string "bubble" appears in FIG. That is, in the present embodiment, the one with the similarity = 0 is not output. Next, the operation of this embodiment will be specifically described with reference to the examples of FIGS.

According to step 11 in FIG. 2, all the similarity lists by the method A are read. In FIG. 3, sets of 30 identifiers and similarities are read. Next, according to step 12 of FIG. 2, first, the first pair in FIG. 4, namely, identifier = NK950215-2322, similarity =
1.00 is read. According to step 13 in FIG.
The one whose identifier is NK950215-2322 is searched from the previously read list. In the case of this embodiment, it can be found first in the list of FIG. Subsequently, according to step 16 of FIG.
805 and 1.00 are synthesized. The above-mentioned s
_In the case where α = 1 and β = 1 using ₀ = α · s ₁ + β · s ₂ , a calculation result of 1.0540805 is obtained, and the similarity in the column of identifier = NK950215-2322 is 1.05.
40805.

Here, for example, if the importance of including the character string "bubble" is low, it is possible to make β smaller than 1. That is, it is possible to perform a search in accordance with the search intention by changing α and β. Thereafter, similarly, identifier = NK950207-008
4. For the identifier = NK950223-2116,
The composite similarity is calculated and replaced. As shown in FIG. 4, since the similarity list of the method B is three, the process is completed three times.

Next, the similarity order rearranging unit 19 in FIG. 1 rearranges the replaced similarity list according to the composite similarity, for example, in descending order. FIG. 5 is an output example of the similarity order rearranging unit 19 of FIG. According to this result, the search result display unit 20 in FIG. 1 retrieves the document from the search target storage unit 11 based on the identifier and displays it to the searcher.

The configuration of the information retrieval apparatus is not limited to the configuration of the embodiment shown in FIG. 1, but each component is constructed by software (program) having the above-described procedure, and the information is stored in a computer system. By executing the program, a general computer system can be used as the information search device of the present invention. FIG. 6 is a block diagram showing an example of a hardware configuration of the computer system. The computer system includes a CPU 100 that executes processing, a memory 101 that stores programs and data, an external storage device 102 that stores programs and data used by the memory 101 or the CPU 100,
It comprises a display 103 for displaying data, a keyboard 104 for inputting data or instructions, and a communication processing device 105 for communicating with other computer systems via a network. The above program is installed in the memory 101 or the external storage device 102 and is
This is executed by U100.

The recording medium of the present invention corresponds to the memory 101 or the external storage device 102. In addition, electronic memory,
Hard disk or floppy disk, CD-R
Portable recording media such as OM and magnetic tape can be used as the recording media of the present invention. By loading the program recorded on the information search program recording medium of the present invention into a general computer system as shown in FIG. 6, it becomes possible to execute the information search method of the present invention on the computer. .

The present invention is not limited to the above embodiment, but can be variously modified and applied within the scope of the claims.

[0035]

As described above, according to the present invention, a plurality of search methods can be used to synthesize a plurality of results, and the ratio of the synthesis can be changed according to the search intention. Can be incorporated into the search result depending on the search intention. Therefore, the problem in the related art is solved, and it is possible to obtain a search result more in line with the search intention.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an information search device according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating an operation of a similarity synthesizing unit in FIG. 1;

FIG. 3 is an example of a result of similarity calculation by the method A according to an embodiment of the present invention.

FIG. 4 is an example of a result of similarity calculation by the method B according to an embodiment of the present invention.

FIG. 5 is an example of a result of similarity synthesis according to one embodiment of the present invention.

FIG. 6 is a configuration diagram of a computer system according to an embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 Search target accumulation part 12 Index part A 13 Index part B 14 Search request input part 15 Similarity calculation part A 16 Similarity calculation part B 17 Similarity combination ratio input part 18 Similarity combination part 19 Similarity order rearrangement part 20 Search result display unit 100 CPU 101 Memory 102 External storage device 103 Display 104 Keyboard 105 Communication processing device

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Tsuneaki Kato 3-19-2 Nishi-Shinjuku, Shinjuku-ku, Tokyo F-term within Nippon Telegraph and Telephone Corporation (reference) 5B075 ND03 NK02 NK32 PP02 PP12 PP25 PQ02 PQ46 PQ75 PR06 QM08 UU06

Claims (12)

[Claims] A plurality of documents are searched for, a similarity between each document in the search target and a word string including one or more words is calculated, and one or a plurality of documents are searched from the search target according to the similarity. In an information search method for presenting a document, a first calculation step of calculating a plurality of types of the similarity for each of the documents by a plurality of methods, and calculating one combined similarity by combining the plurality of types of similarities An information retrieval method, comprising: a second calculation step of performing the following; and a step of presenting the one or more documents from the search target according to the composite similarity. 2. The method according to claim 1, wherein the plurality of methods in the first calculating step include: a method of calculating a similarity by a weight of word association using an associative relationship between each document and the word in the word string; 2. The information retrieval method according to claim 1, wherein the method is a method of calculating a similarity by checking whether or not a character string that matches a character string of a word in the word string exists in a document. Method. 3. The method according to claim 2, wherein the second calculating step includes a step of weighting each similarity calculated in the first calculating step, and summing the weighted similarities. Item 1. The information retrieval method according to item 1. 4. The information retrieval method according to claim 3, wherein the weighting is performed by multiplying the similarity by a coefficient specified from the outside. 5. A search method for a plurality of documents, comprising means for calculating a similarity between each document in the search target and a word string comprising one or more words, wherein one or more documents are searched from the search target according to the similarity. An information search apparatus that presents one or more documents, a first calculating unit that calculates a plurality of types of the similarity for each of the documents by a plurality of units; An information search device comprising: a second calculation unit that calculates a similarity; and a unit that presents the one or more documents from the search target according to the combined similarity. 6. The plurality of means in the first calculating means, comprising: means for calculating a similarity by a weight of word association using an associative relationship between each document and the word in the word string; 6. The information retrieval apparatus according to claim 5, further comprising means for calculating a similarity by examining whether a character string matching the character string of the word in the word string exists in the document. 7. The second calculating means includes means for weighting each similarity calculated by the first calculating means, and means for summing the weighted similarities. The information retrieval apparatus according to claim 5, wherein 8. An information retrieval apparatus according to claim 7, wherein said weighting means comprises: means for externally specifying a coefficient representing said weighting; and means for multiplying said coefficient by said coefficient of similarity. . 9. A plurality of documents to be searched, a similarity between each document in the search object and a word string including one or more words is calculated, and one or a plurality of words are calculated from the search objects according to the similarity. A recording medium storing an information search program for causing a computer to execute a procedure for presenting a document, a first calculation procedure for calculating a plurality of types of the similarity for each document by a plurality of procedures, and the plurality of types of similarities A second calculation procedure of calculating one combined similarity by combining the two, and a procedure of presenting the one or more documents from the search target according to the combined similarity. A recording medium on which an information retrieval program is recorded. 10. The plurality of steps in the first calculation step include: calculating a similarity by weight of word association using each document and the word association in the word string; 10. The information retrieval program according to claim 9, wherein a similarity is calculated by examining whether a character string matching the character string of the word in the word string exists in the document. Recording medium. 11. The second calculation procedure includes a procedure of weighting each similarity calculated by the first calculation procedure, and a procedure of summing up the weighted similarities. A recording medium on which the information retrieval program according to claim 9 is recorded. 12. The computer-readable storage medium according to claim 11, wherein the step of performing the weighting includes a step of externally specifying a coefficient representing the weighting, and a step of multiplying the coefficient by the similarity. Recording medium on which is recorded.