JP3872764B2 - SEARCH DEVICE, SEARCH METHOD, AND PROGRAM - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a search device, a search method, and a program suitable for appropriately searching for a document group in which documents can have a reference relationship.
[0002]
[Prior art]
Conventionally, a document containing a given search term (also referred to as “term”) is extracted from documents published on the Internet or stored in various literature databases, and a score is calculated as appropriate. A search system (also referred to as a “search engine”) that presents documents in the order of the scores is used.
[0003]
On the other hand, in an HTML (HyperText Markup Language) format document used for publishing a document on a website, a link for referring to another HTML format document can be set. Such a reference relationship is not limited to links in HTML format documents, but can be found in various documents.
[0004]
[Problems to be solved by the invention]
Therefore, there is a need for a technique for appropriately searching a document or a document group from a document group having a reference relationship between documents using a given search term, and arranging them appropriately and presenting it to the user. .
[0005]
The present invention solves the above-described problem, and provides a search device, a search method, and a program suitable for appropriately searching a document group in which documents can have a reference relationship. Objective.
[0006]
[Means for Solving the Problems]
To achieve the above object, a search device according to a first aspect of the present invention includes an input reception unit, a storage unit, a search unit, a calculation unit, and an alignment unit, and is configured as follows. .
[0007]
That is, the input receiving unit receives an input of a search term.
[0008]
On the other hand, the storage unit stores documents, terms included therein, and reference relationships between documents.
[0009]
Further, the search unit refers to the storage unit to search for a document including a search term that has received an input, and further includes a document group including a plurality of documents having a reference relationship, and each of the documents included in the document group Searches for items that contain the search term.
[0010]
Then, the calculation unit calculates a score for the document and the document group from the frequency at which the search term appears in the searched document and the frequency at which the search term appears in the searched document group.
[0011]
On the other hand, the sorting unit sorts the searched documents and document groups according to the calculated score, and outputs the results as a search result.
[0012]
A search device according to another aspect of the present invention includes an input reception unit, a storage unit, a search unit, a calculation unit, and an alignment unit, and is configured as follows.
[0013]
That is, the input reception unit receives input of a plurality of search terms.
[0014]
On the other hand, the storage unit stores documents, terms included therein, and reference relationships between documents.
[0015]
Further, the search unit refers to the storage unit to search for a document including at least one of the plurality of search terms that have received the input, and further includes a document group including a plurality of documents having a reference relationship, Each of the documents included in is searched for a document including at least one of the plurality of search terms.
[0016]
Then, the calculation unit calculates the document and the document group from the frequency at which each of the plurality of search terms appears in the searched document and the frequency at which each of the plurality of search terms appears in the searched document group. Calculate the score for.
[0017]
On the other hand, the sorting unit sorts the searched documents and document groups according to the calculated score, and outputs the results as a search result.
[0018]
Further, the above-described search device of the present invention can be configured as follows.
[0019]
That is, the storage unit stores the reference relationship between documents together with their types.
[0020]
On the other hand, the calculation unit calculates the score for the document group from the score calculated for each of the documents and the weight associated with the type of the reference relationship in advance.
[0021]
Further, in the search device of the present invention, the search unit replaces “including at least one of the plurality of search terms” with “the similarity with the plurality of search terms is greater than a predetermined threshold”, and the calculation unit The “similarity between the document and the search term” can be used instead of the “frequency at which the search term appears in the document”. In this case, even a document that does not include any search terms is adopted as a search result if the degree of similarity is high.
[0022]
A search method according to another aspect of the present invention uses a storage unit that stores a document, a term included in the document, and a reference relationship between the documents, an input reception process, a search process, a calculation process, and an alignment. The process is comprised as follows.
[0023]
That is, in the input receiving process, the search term input is received.
[0024]
On the other hand, in the search step, the storage unit is referenced to search for a document including the search term that has received the input, and each of the documents included in the document group includes a plurality of documents having a reference relationship. Searches for items that contain the search term.
[0025]
Further, in the calculation step, a score for the document and the document group is calculated from the frequency at which the search term appears in the searched document and the frequency at which the search term appears in the searched document group.
[0026]
In the alignment step, the searched documents and document groups are aligned according to the calculated scores and output as search results.
[0027]
A search method according to another aspect of the present invention uses a storage unit that stores a document, a term included in the document, and a reference relationship between the documents, an input reception process, a search process, a calculation process, and an alignment. The process is comprised as follows.
[0028]
That is, in the input receiving step, input of a plurality of search terms is received.
[0029]
On the other hand, in the search step, a document including at least one of a plurality of search terms accepted for input is searched with reference to the storage unit, and a document group including a plurality of documents having a reference relationship, Each of the documents included in is searched for a document including at least one of the plurality of search terms.
[0030]
Further, in the calculation step, the document and the document group are calculated from the frequency at which each of the plurality of search terms appears in the searched document and the frequency at which each of the plurality of search terms appears in the searched document group. Calculate the score for.
[0031]
In the alignment step, the searched documents and document groups are aligned according to the calculated scores and output as search results.
[0032]
A program according to another aspect of the present invention is configured to cause a computer to function as the search device described above or to cause the computer to execute the search method described above.
[0033]
The above program can be recorded on various recording media such as a compact disc and a flexible disc, and can be distributed via a computer communication network such as the Internet.
[0034]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the accompanying drawings. The embodiments described below are for explanation, and do not limit the scope of the present invention.
[0035]
Therefore, those skilled in the art can employ embodiments in which each or all of these elements are replaced with equivalent ones, and these embodiments are also included in the scope of the present invention.
[0036]
(Outline configuration of information processing device)
FIG. 1 is a schematic diagram showing a schematic configuration of a typical information processing apparatus that can function as a search apparatus of the present invention. Hereinafter, a description will be given with reference to FIG.
[0037]
The information processing apparatus 101 is controlled by a CPU (Central Processing Unit) 102. When the information processing apparatus 101 is turned on, the CPU 102 executes an IPL (Initial Program Loader) stored in the ROM 103.
[0038]
The IPL reads and executes an OS (Operating System) program stored in a recording medium such as the hard disk 104, the FD attached to the FD drive 110, and the CD-ROM attached to the CD-ROM drive 111. It is a program.
[0039]
After starting the OS, the CPU 102 stores the information in the hard disk 104 or the like in accordance with a user instruction input from the keyboard 105, the mouse 106, or the like, or in accordance with the contents of the setting file described in advance in the hard disk 104 or the like. Run the application program.
[0040]
By executing the application program, the information processing apparatus 101 functions as a search apparatus. The application program is recorded on a storage medium such as the hard disk 104, an FD mounted on the FD drive 110, and a CD-ROM mounted on the CD-ROM drive 111.
[0041]
The CPU 102 uses the RAM 107 as a temporary working storage area when executing these programs. In addition, a register or a cache (not shown) provided in the CPU 102 is used as a temporary working storage area.
[0042]
As the program is executed, the CPU 102 can display the information on a display device 108 such as a liquid crystal display or a CRT (Cathode Ray Tube) in order to report the result to the user or to show the progress of the program. In the instruction operation using the mouse 106, the cursor displayed on the screen is moved by moving the mouse 106, and the menu item pointed to by the cursor can be selected by clicking the mouse 106.
[0043]
The information processing apparatus 101 can communicate with a computer communication network such as the Internet via an interface 109 such as a NIC (Network Interface Card) or a modem. The document data received via the interface 109 can be processed, the processing result can be transmitted via the interface 109, the program received via the interface 109 can be executed, and the like.
[0044]
FIG. 2 is a schematic diagram illustrating a schematic configuration of a search device according to one embodiment of the present invention. FIG. 3 is a flowchart showing a flow of control of processing executed by the search device of the present embodiment. Hereinafter, description will be given with reference to these drawings.
[0045]
The search device 201 includes an input receiving unit 202, a storage unit 204, a search unit 203, a calculation unit 205, and an alignment unit 206.
[0046]
First, the input receiving unit 202 receives an input of a search term (step S301). The following various techniques can be applied to the search term input method.
(A) A word is input and this is used as a search term.
(B) A natural sentence is input, a morphological analysis is performed to obtain a word, and this is used as a search term.
(C) In (a) and (b) above, further normalization of words is performed to obtain an appropriate search term.
[0047]
Accordingly, the keyboard 105 and the mouse 106 work together with the CPU 102 and the display device 108 to function as the input receiving unit 202.
[0048]
On the other hand, the storage unit 204 stores documents, terms included therein, and reference relationships between documents. In the present embodiment, for example, a document itself in the Internet may be used as it is, or the document itself may be stored in the storage unit 204. In the former case, for example, the following data may be stored as one record.
(1) A URL (Universal Resource Locator) where a document is located in the Internet.
(2) Search terms included in the document and their appearance frequency. If the search term is normalized as described above, the identification number assigned to the search term may be stored instead of storing the spelling of the search term itself.
(3) URL of other document referred to by the document and information on the type of reference relationship. If each record is managed by a record number, the record number of the document may be used instead of the URL.
[0049]
FIG. 4 is an explanatory diagram showing a reference relationship structure between three documents.
[0050]
Document A is a document about “structured search”, in which document B “What is search” is cited, and document C “what is the effect?” Is referred to as a document that explains document A in more detail. Yes.
[0051]
Therefore, in the record for document A, information such as “citation” is stored as the reference relationship type of “document B”, and “explanation” is stored as the reference relationship type of “document C”. It will be.
[0052]
Therefore, the hard disk 104 functions as the storage unit 204.
[0053]
Further, the search unit 203 refers to the storage unit 204 and searches for a document including the search term that has received the input (step S302). That is, a record including the search term is first obtained from the record stored as described above.
[0054]
Further, a document group composed of a plurality of documents having a reference relationship with the document (reference source) searched in step S302 is searched for , each of which includes the search term (step S303). . That is, for each record (reference source) obtained in step S302, a record (reference destination) having a reference relationship with the record (reference source) is obtained, and the obtained record (reference destination) having the reference relationship is obtained. There determines whether to include the search term, and "document group" one collectively record (reference) containing the search term.
[0055]
That is, the search result document (reference source) refers to a document (reference destination) included in the document group.
[0056]
Therefore, the CPU 102 functions as the search unit 203 in cooperation with the hard disk 104.
[0057]
Then, the calculation unit 205 calculates a score for the document from the searched document and the search term, and calculates the score for the document group from the score calculated for each of the documents included in the searched document group. A score is calculated (step S304).
[0058]
In the present embodiment, the search terms included in each document and the frequency thereof are stored in the storage unit 204. Therefore, it is preferable to use the vector inner product method for calculating the score between each document and the search terms. is there.
[0059]
Then, the score of each document is obtained by multiplying the obtained score of each document by a weight previously associated with the type of the reference relationship to obtain a total. For example, in the case of the “citation” relationship in the above example, the document B (reference destination) is viewed more heavily than the document A (reference source), and in the case of the “explanation” relationship, the document A (reference source). It is possible to adopt a technique such that the document is viewed heavier than the document C (reference destination).
[0060]
Other calculation methods for the score of the document group will be described later.
[0061]
Therefore, the CPU 102 functions as the calculation unit 205 in cooperation with the hard disk 104.
[0062]
Finally, the aligning unit 206 aligns the retrieved documents and document groups according to the calculated score, and outputs them as a retrieval result (step S305).
[0063]
Therefore, the CPU 102 functions as the alignment unit 206 in cooperation with the display device 108 (when displaying the search result) and the hard disk 104 (when storing the search result).
[0064]
As described above, when only one search term input is accepted, all the documents included in the search result document group include the search term. On the other hand, when receiving a plurality of search term inputs, the document searched in step S302 (reference source) and the document to which it refers (reference destination) includes at least one of the plurality of search terms. It is desirable to calculate the score.
[0065]
(Other score calculation methods)
In the above embodiment, the score calculation for the search result document (reference source) and the score calculation for the document (reference destination) included in the document group of the search result are performed independently, and the score obtained for these is calculated. The score of “searched document and document group” is obtained by appropriately weighting according to the type of the reference relationship.
[0066]
In the following, various methods for calculating other scores will be described.
[0067]
First, consider one document (hereinafter referred to as “total document”) in which a retrieved document and a document included in the retrieved document group are linked. Then, the score is calculated based on the frequency with which the search term appears in the total document. If the search term appearance frequency in the total document is the sum of the search term appearance frequencies in the searched document and the documents included in the searched document group, the score of this method can be easily obtained. Can be obtained. This process will be described below in accordance with the flowchart shown in FIG. That is, an input of a search term is accepted (step S301), a document including the search term (“reference source”) is searched (step S302), and a document group that is referred to by the reference source and includes the search term. (Step S303) is the same as in the above embodiment, but when the score is calculated (Step S304), the reference source document and the document referred to by the reference source include the search term. The score of the total document which connected the document group is calculated. Then, the total documents are arranged and output in the order of the scores (step S305).
[0068]
In addition to this, for each search term, there is also a method of considering a maximum value of the appearance frequency of the term in the searched document and the documents included in the searched document group, and obtaining a score by this maximum value. is there.
[0069]
Furthermore, in the above embodiment, each document necessarily includes a search term. However, when the Vector Space Model is adopted as a method for searching for a document, the similarity between the search term and the document is calculated as an inner product. Calculate by etc. In this case, even a document with a high degree of similarity may not include any search terms. Therefore, in this case, the condition “including at least one search term” is changed to a condition “similarity to the search term is greater than a certain threshold”, and changed to “frequency at which each search term appears”. Then, “similarity to each search term” may be adopted.
[0070]
The storage unit 204 stores the document, the terms included in the document, and the reference relationship between the documents. When these pieces of information are registered in the storage unit 204, the term included in a certain document is stored. "A term that can be extracted from a document that is not included in the document but has a reference relationship with the document and that corresponds to the reference relationship" may be added.
[0071]
For example, a document that defines a frame in an HTML document includes only link information with respect to other documents, and therefore, it is general that almost nothing corresponding to a term is included. Therefore, when registering the document defining the frame in the storage unit 204, the terms included in the document referenced from the frame are registered as the terms of the document defining the frame.
[0072]
By doing this, it is possible to add a role as an index to a document defining an HTML frame, and to make the document a representative (upper) page of a document group, and the search result is good. .
[0073]
【The invention's effect】
As described above, according to the present invention, a search device, a search method, and a computer that can suitably search for a document group in which documents can have a reference relationship are realized by a computer. A program can be provided.
[Brief description of the drawings]
FIG. 1 is a schematic diagram showing a schematic configuration of a typical information processing apparatus in which a search device of the present invention is realized.
FIG. 2 is a schematic diagram showing a schematic configuration of a search device according to one embodiment of the present invention.
FIG. 3 is a flowchart showing a flow of control of processing executed by the search device.
FIG. 4 is an explanatory diagram showing a reference relationship structure between three documents.
[Explanation of symbols]
101 Information processing apparatus 102 CPU
103 ROM
104 HD
105 Keyboard 106 Mouse 107 RAM
108 display device 109 interface 110 FD drive 111 CD-ROM drive 201 search device 202 input reception unit 203 search unit 204 storage unit 205 calculation unit 206 alignment unit

Claims (7)

An input receiving unit for receiving search terms,
A storage unit for storing a document, a term included in the document, and a reference relationship between documents;
A document group including one or more documents referred to from the retrieved document is retrieved by referring to the storage unit and including a search term that has received the input, and is included in the document group A search unit for searching for a document group in which each of the documents includes the search term;
A calculation unit for calculating a score for the total document from the frequency of occurrence of the search term in the total document obtained by connecting the searched document and the document included in the searched document group;
A search apparatus comprising: an alignment unit that aligns the total document according to the calculated score and outputs the result as a search result. An input accepting unit that accepts input of multiple search terms;
A storage unit for storing a document, a term included in the document, and a reference relationship between documents;
A document group including one or more documents referred to from the retrieved document is searched by referring to the storage unit and searching for a document including at least one of the plurality of search terms that have received the input. A search unit for searching for a document group in which each of the documents included therein includes at least one of the plurality of search terms,
A calculation unit for calculating a score for the total document from a frequency at which each of the plurality of search terms appears in a total document obtained by connecting the searched document and the document included in the searched document group;
A search apparatus comprising: an alignment unit that aligns the total document according to the calculated score and outputs the result as a search result. The search device according to claim 1,
Instead of “searching for a document including a search term”, the search unit “searches for a document whose similarity to the search term is greater than a predetermined threshold”, and “each of the documents includes the search term” Instead of “searching for a group”, “search for a document group in which the similarity between each of the documents and the search term is greater than a predetermined threshold”,
The calculation unit calculates the score from the “similarity between the total document and the search term” instead of calculating the score from the “frequency of appearance of the search term in the total document”. Search device to do. The search device according to claim 2,
Instead of “searching for a document including at least one of a plurality of search terms”, the search unit “searches for a document having at least one similarity with each of the plurality of search terms greater than a predetermined threshold”. However, instead of “searching for a document group in which each document includes at least one of the plurality of search terms”, at least one of the similarities between each of the documents and each of the plurality of search terms is predetermined. Search for documents larger than the threshold of "
The calculation unit calculates the score from “similarity between the total document and each of the plurality of search terms” instead of calculating the score from “frequency of occurrence of the search term in the total document”. A search device characterized by that. Executed by a search device including an input reception unit, a document, a term included therein, a reference relationship between documents, a storage unit, a search unit, a calculation unit, and an alignment unit. A search method,
  An input receiving step in which the input receiving unit receives an input of a search term;
  The search unit refers to the storage unit to search for a document including a search term that has received the input, and further includes a document group including one or more documents referred to from the searched document. A search step of searching for a document group in which each of the documents included therein includes the search term;
  A calculation step in which the calculation unit calculates a score for the total document from a frequency at which the search term appears in a total document obtained by connecting the searched document and a document included in the searched document group;
  An alignment step in which the alignment unit aligns the total document according to the calculated score and outputs it as a search result.
  A search method comprising: Executed by a search device including an input reception unit, a document, a term included therein, a reference relationship between documents, a storage unit, a search unit, a calculation unit, and an alignment unit. A search method,
  An input receiving step in which the input receiving unit receives input of a plurality of search terms;
  The search unit refers to the storage unit to search for a document including at least one of a plurality of search terms that have received the input, and further, from one or more documents referenced from the searched document A search step of searching for a document group in which each of the documents included therein includes at least one of the plurality of search terms,
  A calculation in which the calculation unit calculates a score for the total document from the frequency at which each of the plurality of search terms appears in the total document obtained by connecting the searched document and the document included in the searched document group. Process,
  An alignment step in which the alignment unit aligns the total document according to the calculated score and outputs it as a search result.
  A search method comprising: A program that causes a computer to function as each unit of the search device according to any one of claims 1 to 4.

Images