JP3411246B2 - Query output device, document search system, query output method, document search method, and recording medium - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention can specify a document to be searched by a sentence rather than a keyword and is suitable for generating a query suitable for a search from the sentence, a query output device, a query output method, and a document. The present invention relates to a search system, a document search method, and a computer-readable recording medium recording a program for realizing these.

[0002]

2. Description of the Related Art With the development of WWW (World Wide Web) technology, a document search technology for searching a desired document from a large number of documents is becoming more and more important.

Conventionally, the following technique has been used in such document retrieval. That is, the user specifies a keyword for searching for a desired document, or inputs a query that specifies a condition for searching for a document in more detail.

Then, the document search engine obtains the number of times the keyword is included in the document, the degree of matching between the condition specified by the query by the vector space method and the document,
The score of the document is calculated based on these.

Finally, the documents with high scores are presented to the user in order as search results.

In the vector space method, a method of appropriately weighting a keyword for retrieval (hereinafter referred to as a "term") included in a query, or a document included in a search range of a document search engine, Each term extracted from the document (hereinafter referred to as "search term")
There is also proposed a method for accelerating score calculation by separately storing.

[0007]

However, in the method of inputting a keyword, when a plurality of keywords are input, whether one wants to search for one that includes all the keywords or one that includes any of the keywords, There is a possibility that the search result will be clear and that the search result that the user does not expect is obtained.

Further, if the designated keyword is not a prepared search term, there is a possibility that a significant result will not be obtained as a search result.

In addition, it is difficult for a general user to compose a query for a document search engine by himself or to adjust the weight of terms included in the query.

Further, there is a demand for a method of automatically generating an appropriate query from a document when a user specifies a document to be searched with a natural sentence.

The problem to be solved by the present invention is to specify a document to be searched by a sentence rather than a keyword, generate a query suitable for the search from the sentence, and further search the document by the query. .

[0012]

In order to solve the above problems, the query output device according to the first aspect of the present invention has a score obtained by applying a query including a term to a document within a predetermined range. The query given to the document search engine that acquires the documents contained in is output.

The query output device includes a character string input receiving unit, a term extracting unit, an input control unit, a query generating unit,
It is configured to include a message display unit, a correction input reception unit, a correction control unit, and a query output unit.

Here, the character string input receiving unit receives the input of the character string.

The term extractor morphologically analyzes the given character string and extracts the terms contained therein.

When the input of the character string is accepted by the character string input accepting unit, the input control unit gives the character string to the term extracting unit.

The query generation unit generates a query including the terms extracted by the term extraction unit.

When the message display condition is satisfied, the message display unit displays a message previously associated with the message display condition.

The correction input receiving unit receives a correction input for correcting the character string input by the character string input receiving unit when the message is displayed by the message display unit.

When the correction input is received by the correction input receiving unit, the correction control unit supplies the term extracting unit with the character string corrected by the correction input.

The query output unit outputs the query generated by the query generation unit when the query output condition is satisfied.

In the above-mentioned query output device, when the user specifies a document to be searched with a natural sentence, a query suitable for the search is generated from the sentence. Further, for example, when there is a possibility that a significant search result may not be obtained, the user is notified of that fact and requested to input further detailed search conditions. Therefore, the user can obtain a desired search result simply by inputting a natural sentence.

Further, the query output device can be configured to further include a term storage unit.

Here, the term storage unit stores a plurality of terms.

On the other hand, when the terms included in the query generated by the query generation unit do not match any of the plurality of terms stored in the term storage unit, the message display condition is satisfied.

In the above query output device, if the query generated from the sentence input by the user includes a term that is not a search term, a significant search result may not be obtained. Therefore, the user is requested to correct the search condition. Urge. Therefore, the user can specify more detailed conditions before executing the inquiry to the document search engine.

Further, the query output device can be configured so as to further include a parameter input acceptance unit.

Here, the parameter input receiving unit receives the input of the generation parameter given to the query generating unit.

On the other hand, the query generated by the query generation unit includes, in addition to terms, weights associated with each of the terms and referred to when obtaining the score.

Further, the query generation unit determines the weight of each of the terms extracted by the term extraction unit based on the generation parameter input by the parameter input reception unit, and associates the term with the term. And a weight determined by the above are generated.

In the above-mentioned query output device, the user can input parameters for adjusting the weight of terms included in the query. For example, by changing the weight given to each term depending on whether the parameter is "loose search" or "tight search", the user's request It is possible to generate an appropriate query according to the application and usage.

Further, the query output device can be configured so as to further include a parameter input acceptance unit.

Here, the parameter input receiving unit receives the input of the generation parameter given to the query generating unit.

On the other hand, the query generated by the query generating unit includes, in addition to the term, essential term information indicating whether or not the documents acquired by the document search engine are limited to those including the term.

Further, the query generation unit determines the essential term information of each of the terms extracted by the term extraction unit according to the generation parameter input of which is received by the parameter input reception unit.
A query including the required term information determined in association with this is generated.

In the above query output device, the user needs to set a parameter for adjusting whether or not the term included in the query needs to appear in the document (in this case, the term is referred to as "essential term"). You can enter. For example, depending on whether the parameter is "loose search" or "tight search", one of the terms or
By making all required terms, it is possible to generate an appropriate query according to the user's request and usage.

In order to solve the above problems, a document retrieval system according to a second aspect of the present invention is configured to include the above query output device, a document retrieval engine section, and a document output section.

Here, the document search engine unit acquires a document in which the score obtained by applying the query output by the query output device to the document falls within a predetermined range.

On the other hand, the document output unit outputs the documents acquired by the document output unit and the document search engine unit.

In the document search system, the query generated by the query output device is given to the document search engine to perform an actual search. Although the processing in the document search engine may take a long time, in the above document search system, before the processing in the document search engine, a query expected to obtain an appropriate search result by the query output device is previously After creating, make an inquiry with the document search engine. Therefore, the number of inquiries required by the user for the document search can be made extremely small.

In order to solve the above problems, a query output method according to a third aspect of the present invention obtains a document in which a score obtained by applying a query including a term to a document falls within a predetermined range. The query given to the document search engine is output, and a character string input receiving step, a term extracting step, an input controlling step, a query generating step, a message displaying step, a correction input receiving step, a correction controlling step, and a query. And an output step.

Here, in the character string input receiving step, the input of the character string is received.

In the term extracting step, the given character string is subjected to morphological analysis, and terms included in this are extracted.

In the input control step, when the input of the character string is accepted in the character string input accepting step, the term extracting step causes the term to be extracted from the character string.

In the query generation step, a query including the terms extracted in the term extraction step is generated.

In the message display step, when the message display condition is satisfied, the message previously associated with the message display condition is displayed.

In the correction input receiving step, when the message is displayed in the message displaying step, the correction input for correcting the character string input in the character string input receiving step is received.

In the correction control step, when the correction input is received in the correction input receiving step, the term extracting step causes the term to be extracted from the character string corrected by the correction input.

In the query output step, if the query output condition is satisfied, the query generated in the query generation step is output.

In the above query output method, if the term included in the query generated in the query generating step does not match any of a plurality of prestored terms, the message display condition is satisfied. Can be configured.

The query output method can be configured so as to further include a parameter input receiving step.

Here, in the parameter input receiving step, the input of the generation parameter referred to in the query generating step is received.

On the other hand, the query generated in the query generating step includes, in addition to the terms, weights associated with each of the terms and referred to when obtaining the score.

Further, in the query generation step, the weight of each of the terms extracted in the term extraction step is determined by the generation parameter input in the parameter input reception step, and the term and A weight including the weight determined in association with the query is generated.

The query output method can be configured to further include a parameter input receiving step.

Here, in the parameter input receiving step, the input of the generation parameter referred to in the query generating step is received.

On the other hand, the query generated in the query generating step includes, in addition to the term, essential term information indicating whether or not the documents acquired by the document search engine are limited to those including the term.

Further, in the query generation step, the essential term information of each term extracted in the term extraction step is determined by the generation parameter input of which is accepted in the parameter input acceptance step,
A query including the term and the essential term information determined in association with the term is generated.

In order to solve the above problems, the document search method of the present invention is configured to include a document search step and a document output step.

Here, in the document search step, a document in which the score obtained by applying the query output by the query output method to the document is included in a predetermined range is acquired.

In the document output step, the document acquired in the document search step is output.

In order to solve the above-mentioned problems, the computer-readable recording medium of the present invention has a computer as a character string input receiving unit, a term extracting unit, an input control unit, a query generating unit, a message display unit, a correction input receiving unit. , Correction controller,
Also, it is configured to record a program that functions as a query output unit.

Here, the character string input receiving unit receives the input of the character string.

The term extracting unit morphologically analyzes the given character string and extracts terms contained in this.

When the input of the character string is accepted by the character string input accepting unit, the input control unit gives the character string to the term extracting unit.

The query generation unit generates a query including the terms extracted by the term extraction unit.

When the message display condition is satisfied, the message display section displays the message previously associated with the message display condition.

The correction input receiving unit receives a correction input for correcting the character string input by the character string input receiving unit when the message is displayed by the message display unit.

When the correction input is received by the correction input receiving unit, the correction control unit supplies the term extracting unit with the character string corrected by the correction input.

The query output section outputs the query generated by the query generation section when the query output condition is satisfied.

Further, in the above computer-readable recording medium, the program causes the computer to
It can be configured to further function as a term storage unit.

Here, the term storage unit stores a plurality of terms.

On the other hand, when the terms included in the query generated by the query generation unit do not match any of the plurality of terms stored in the term storage unit, the message display condition is satisfied.

Further, in the above computer-readable recording medium, the program causes the computer to
It can be configured to further function as a parameter input reception unit.

Here, the parameter input receiving unit receives the input of the generation parameter given to the query generating unit.

On the other hand, the query generated by the query generating section includes, in addition to terms, weights associated with each of the terms and referred to when obtaining the score.

Further, the query generation unit determines the weight of each of the terms extracted by the term extraction unit based on the generation parameter input by the parameter input reception unit, and associates the term with the term. And a weight determined by the above are generated.

Further, in the above computer-readable recording medium, the program causes the computer to
It can be configured to further function as a parameter input reception unit.

Here, the parameter input reception unit receives the input of the generation parameter given to the query generation unit.

On the other hand, the query generated by the query generation unit includes, in addition to the term, essential term information indicating whether or not the documents acquired by the document search engine are limited to those including the term.

Further, the query generation unit determines the essential term information of each of the terms extracted by the term extraction unit according to the generation parameter input of which is accepted by the parameter input acceptance unit.
A query including the required term information determined in association with this is generated.

Further, in the above computer-readable recording medium, the program causes the computer to
It can be configured to further function as a document search engine unit and a document output unit.

Here, the document search engine unit acquires a document whose score obtained by applying the query output by the query output unit to the document falls within a predetermined range.

On the other hand, the document output unit outputs the document acquired by the document search engine unit.

[0085]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the accompanying drawings.

(Outline Configuration) FIG. 1 shows a typical information processing apparatus (including a general-purpose computer, various terminals, mobile terminals, mobile phones, game machines, and other dedicated equipment) that functions as the query output apparatus of the present invention. It is a schematic diagram which shows a schematic structure. Hereinafter, description will be given with reference to FIG.

The information processing apparatus 101 has a CPU (Central
Processing Unit (Central Processing Unit) 102. When the information processing apparatus 101 is powered on,
The CPU 102 has an IPL (Initial Program) stored in a ROM (Read Only Memory) 103.
Loader; initial program loader) is executed.

The IPL is the hard disk 104, FD
(Floppy Disk) drive 110
FD, CD-ROM (Compact Disk ROM) installed in
It is a program for reading and executing an OS (Operating System) program stored in a recording medium such as a CD-ROM mounted in the drive 111.

After booting the OS, the CPU 102 is stored in the hard disk or the like according to the user's instruction input by the keyboard 105 or the mouse 106 or according to the contents of the setting file previously described in the hard disk or the like. Run the application program.

In a small-scale information processing device such as a mobile terminal, an embodiment in which the IPL itself functions as an OS and an application program can be adopted.

The CPU 102 uses a RAM (Random Access Memory) 107 as a temporary work storage area when executing a program. In addition, a register and a cache (not shown) provided in the CPU 102 are used as a temporary work storage area.

In order to report the result to the user or show the progress along with the execution of the program, 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 the instruction operation by the mouse 106, the cursor displayed on the screen is moved by moving the mouse 106, and the menu item pointed by the cursor can be selected by clicking the mouse 106.

The information processing apparatus 101 has a NIC (Network
Interface card, interface such as modem 10
9 can communicate with a computer communication network such as the Internet. It is possible to send a query to another server computer (literature search engine) via the interface 109 and receive the search result sent as a response from this. It is also possible to execute the program received via the interface 109.

FIG. 2 is a schematic diagram showing a schematic configuration of each function when the information processing apparatus 101 functions as the query output apparatus of the present invention.

The query output device 201 includes a character string input acceptance unit 202, a term extraction unit 203, and an input control unit 204.
A query generation unit 205 and a message display unit 206
A correction input reception unit 207, a correction control unit 208, and a query output unit 209.

Further, the term storage unit 210 and the parameter input reception unit 211 can be provided. Hereinafter, description will be given based on a configuration including these.

The character string input receiving unit 202 receives an input of a character string representing a natural sentence for retrieval from the user. Therefore, the input device such as the keyboard 105 and the mouse 106 functions as the character string input receiving unit 202.

The term extracting unit 203 morphologically analyzes the given character string and extracts terms included in this. Therefore, the CPU 102 functions as the term extracting unit 203.

The input control unit 204 includes the character string input receiving unit 2
When the input of the character string is accepted by 02, the character string is given to the term extracting unit 203. Therefore, CP
The U 102 works in cooperation with an input device such as the keyboard 105 and the mouse 106, and functions as the term extraction unit 203.

The query generation unit 205 includes the term extraction unit 20.
Generate a query that includes the terms extracted by 3. Therefore, the CPU 102 functions as the query generation unit 205.

When the message display condition is satisfied, the message display unit 206 displays a message previously associated with the message display condition. Therefore, the display device 108 functions as the message display unit 206 in cooperation with the CPU 102.

When the message display unit 206 displays a message, the correction input receiving unit 207 receives a correction input for correcting the character string input by the character string input receiving unit 202. Therefore, the input device such as the keyboard 105 or the mouse 106 functions as the correction input receiving unit 207.

The correction control unit 208 includes the correction input reception unit 20.
When the correction input is accepted by 7, the character string corrected by the correction input is given to the term extracting unit 203. Therefore, the CPU 102 functions as the correction control unit 208.

When there is a correction input, the term extraction unit 2
03 is driven again. In the present embodiment, the message displayed by the message display unit 206 indicates that “a significant search result may not be obtained.” Therefore, when the message is not displayed, the query output condition is satisfied.

When the query output condition is satisfied, the query output unit 209 outputs the query generated by the query generation unit 205 and transmits the query to the document search server. Therefore, the interface 109 functions as the query output unit 209.

The term storage unit 210 stores a plurality of terms. Therefore, the ROM 103 and the hard disk 1
04, the FD installed in the FDD 110, the CD-ROM installed in the CD-ROM drive 111, and another computer connected via the interface 109 function as the term storage unit 210.

In this embodiment, the terms included in the query generated by the query generating section 205 are the term storage section 2
If there is no match with any of the multiple terms stored in 10, there is a possibility that significant search results may not be obtained.
The message display condition is satisfied.

The parameter input acceptance unit 211 accepts the input of the generation parameter given to the query generation unit 205. Therefore, the input device such as the keyboard 105 and the mouse 106 functions as the parameter input reception unit 211.

In the present embodiment, the query generated by the query generating unit 205 includes, in addition to terms, weights associated with each of the terms and referred to when obtaining the score.

Further, the query generation unit 205 determines the weight of each of the terms extracted by the term extraction unit 203 based on the generation parameter input by the parameter input reception unit 211, and associates it with the term. And a weight determined by the above are generated.
The method for determining the weight will be described later.

Further, the query generated by the query generation unit 205 includes, in addition to the term, essential term information indicating whether or not the documents acquired by the document search engine are limited to those including the term.

The query generation unit 205 uses the term extraction unit 20.
The essential term information of each of the terms extracted by 3 is determined by the generation parameter input of which is accepted by the parameter input accepting unit 211, and the term and the essential term information determined in association with the term are included. Generate a query. The generation of the essential term information will be described later.

Besides, it is possible to configure a document search engine unit (not shown) and a document output unit (not shown) on the same information processing device 101 as the query output device 201. However, from the viewpoint of load distribution and centralized management of the literature database, it is desirable to separate them.

In addition, the hard disk 104, the FD installed in the FD drive 110, the CD-ROM installed in the CD-ROM drive 111, the ROM 103, and the RA.
M107 and the equivalent of these readable information storage devices of other computers connected via the interface 109 function as the recording medium of the present invention by recording the program executed by the CPU 102.

As described above, the information processing device 1 shown in FIG.
01 functions as the query output device 201 of the present invention shown in FIG. This is an exemplary implementation of the query output device 201 of the present invention.

On the other hand, in a portable terminal or the like, a nonvolatile semiconductor memory (flash memory card) is used as an external storage device.
A battery-backed RAM can be used. In this case, the hard disk 104,
The FD drive 110, the CD-ROM drive 111, etc. are not necessary.

(Processing Flow) FIG. 3 is a flowchart showing a processing procedure when the information processing apparatus 101 shown in FIG. 1 functions as the query output apparatus 201 of the present invention.

First, the CPU 102 accepts an input of a sentence indicating a search condition for specifying a document to be searched from the user (step S301). The user can input a natural sentence such as “Japanese-Chinese joint venture” as a search condition.

Next, the CPU 102 morphologically analyzes the input character string and extracts terms included in it (step S302). In the case of the above input example, the result of the term morphological analysis is morphologically analyzed as "Japan / and / China / of / joint venture / company". Here, / indicates a morpheme delimiter. From here, the terms "Japan" and "China"
Extract "joint venture / company".

If the term includes other terms, these terms are also extracted (step S303). Terms included in other terms are called "subterms". For example, the sub-terms of “joint venture / company” are “joint venture” and “company”.

Further, related terms such as synonyms and synonyms of these terms and sub-terms (hereinafter referred to as “related terms”)
Say. ) Is extracted (step S304).

Information on a certain term and related terms is preferably stored in the hard disk 104 or the like in advance. For example, as related terms of "Japan", "Japan", "Nippon", "Japan", and "Japan"
In addition to “n”, “East Asia”, “Hokkaido”, “Kyushu”, “Shikoku”, etc. are extracted.

On the other hand, the CPU 102 acquires the search parameters separately input by the user (step S30).
5). This search parameter is, for example, step S3.
It is possible to adopt an embodiment in which the data are simultaneously input at 01. FIG. 4 is a display example of a search condition input form in which a user inputs a search condition character string and a search parameter.

The text field 402 of the search condition input form 401 receives input of a search condition character string. The slider 403 receives input of search parameters. The user operates the mouse 106 and the keyboard 105 to input respective information in these forms.

In the present embodiment, when the user clicks the query output button 404, the processing after step S301 is started.

Next, the CPU 102 checks whether or not the extracted terms include a term that is not a predetermined term (step S306). The “predetermined term” is the hard disk 104 or the CD-RO.
It is stored in M or the like. These are search keywords that can be used in the document search engine to which the query output device 201 is connected. If there is a term that is not a search keyword (step S306; Yes), a message that “significant search results may not be obtained” is displayed (step S307), and the user is prompted to re-enter search conditions.

In the present embodiment, the message display condition is satisfied when "there is a term that is not a search keyword", but the message display condition can be changed as appropriate.

FIG. 5 is a display example showing how this message is displayed. The message display area 501 displays the message in the message display area 5
02, and a re-entry button 503 to be clicked when re-entering, and a forcible button 50 to be clicked when forcing, that is, when performing a search with the search conditions as they are.
4 and are displayed.

The CPU 102 determines whether or not the user re-inputs (step S308), and when re-inputting (step S308; Yes), returns to step S301.

On the other hand, when all the terms are search keywords (step S306; No) and when the user selects the forced line (step S308; No), CP
U102 produces | generates a query from a term, the part-of-speech information of the said term acquired by morphological analysis, the information of dependency, and a search parameter (step S309).

The query generated here includes the following information. ·term. Includes sub-terms and related terms. -The weight of the term. It is used when calculating the scores of documents by the vector space method.・ Required term information. Indicates whether the term is an essential term.

The method of determining the weight of the term and the essential term information will be described later.

Finally, the CPU 102 outputs the generated query (step S310). As the output destination,
For example, the RAM 107 can be adopted. The output query is used to make an inquiry to the document search engine, as described later.

FIG. 6 is a flow chart showing the flow of processing for making an inquiry to the document search engine using the output query.

The CPU 102 transmits the query output to the RAM 107 in step S310 to the document search engine server via the interface 109 (step S601).

The document search engine server receives the query (step S602), and based on the query,
The score of each document is calculated (step S603), a predetermined number of high-ranking documents are acquired (step S604), and these are transmitted as search results (step S605).

The CPU 102 has an interface 109.
The document of the search result is received via (step S60
6) Then, the document is displayed on the display device 108 (step S607), and this processing ends.

As described above, the user can specify a search condition by inputting a natural sentence, but if there is a possibility that a significant search result may not be obtained, that effect is displayed. By prompting the user to re-input, desired search results can be easily obtained, and the number of inquiries to the document search engine can be reduced.

(Determination of Weight) In the query generation unit 205 and step S309, a query is generated from a term, part-of-speech information and dependency information of the term obtained by morphological analysis, and a search parameter. To do.
Hereinafter, an example of a method for determining the term weight in this query will be described.

In Japanese, since the relation between nouns and compound nouns are often used, it is necessary to consider them also in the document search. Therefore, these are categorized as follows from the tendency of the search result, and the weight of each term is determined.

The first type is a concatenation of general nouns and general nouns (including a case where particles are continuous with a particle "no" in between. The same applies hereinafter).
This is the case. In this case, the common noun placed before is often the subject. For example, "meal / therapy"
In the case of “meal / no / therapy”, “golf / tournament”, and “golf / no / tournament”, the main subjects are “meal” and “golf”, respectively. Therefore, the weight of terms placed before these subjects are increased.

The second type is the case of terms containing subterms. When the meanings of subterms and terms are significantly different, the weight of the subterm is lowered and the weight of the entire term is raised. For example, "Regional / Bank""Job /
In the “Agreement,” the overall weight of these is increased, and the weight of “regional,” “bank,” “employment,” and “agreement” is lowered.

In the present embodiment, these weight determination methods are adopted, but heuristics for determining other weights can be adopted according to the results of research by natural language. Further, similar processing can be performed for other languages.

(Determination of Required Term Information) Query Generation Unit 2
In 05 and step S309, a query is generated from the term, the part-of-speech information and the dependency information of the term obtained by the morphological analysis, and the search parameter. Below, an example of a method of determining the essential term information in this query will be described.

Whether or not a required term is related to whether or not the term is a subject. Therefore, in the above type of weight determination, the term “increasing weight” can be made an essential term.

In the second type, the query may be created without using the sub-terms “region”, “bank”, “employment”, and “agreement”.

(Adjustment by Search Parameter) In this embodiment, the term weight and the essential term information are determined according to the search parameter input by the user. In the search condition input form 401 shown in FIG. 4, a slider 403 can be used to input search parameters from "tight search" to "soft search" as values in several stages.

In the case of "tight search", the difference in height between weights is increased, or the proportion of essential terms is increased. on the other hand,
In the case of "loose search", reduce the difference in weight level,
Alternatively, reduce the proportion of mandatory terms.

For example, for “cheap / travel” (when the term placed later is the subject), “loose search”
The information included in the query is listed in order from to "tight search".

"Cheap" is light weight, "Travel" is light weight

"Cheap" has a small weight, and "Travel" has a large weight

"Cheap" has a small weight, and "travel" has a large weight and is essential

"Cheap" is heavy and "travel" is heavy and essential

"Cheap" is heavy and essential, and "Travel" is heavy and essential

The relationship between these search parameters and the generated query can be modified as appropriate, and modified embodiments are also included in the scope of the present invention.

[0156]

According to the present invention, a document to be searched can be designated by a sentence rather than a keyword, and a query output device, a query output method, and a document search system suitable for generating a query suitable for the search from the sentence. It is possible to provide a document search method and a computer-readable recording medium in which a program for realizing these is recorded.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing a schematic configuration of a typical information processing device that functions as a query output device of the present invention.

FIG. 2 is a schematic diagram showing a schematic configuration of each function of the query output device of the present invention.

FIG. 3 is a flowchart showing a processing procedure of a query output device of the present invention.

FIG. 4 is an explanatory diagram showing a display example of a search condition input form.

FIG. 5 is an explanatory diagram showing a display example of a message display form.

FIG. 6 is a flowchart showing a procedure of processing for performing a document search using a query output by the query output device of the present invention.

[Explanation of symbols]

101 Information processing device 102 CPU 103 ROM 104 hard disk 105 keyboard 106 mice 107 RAM 108 display device 109 interface 110 FD drive 111 CD-ROM drive 201 query output device 202 Character string input reception part 203 Term Extraction Unit 204 Input control unit 205 Query generator 206 Message display area 207 Correction input acceptance section 208 Correction control unit 209 Query output section 210 term storage 211 Parameter input acceptance section 401 Search condition input form 402 text field 403 slider 404 Query output button 501 message display form 502 Message display area 503 Re-enter button 504 push button

Continuation of the front page (56) Reference JP-A-3-179563 (JP, A) JP-A-2-82354 (JP, A) JP-A-2-176868 (JP, A) JP-A-5-324728 (JP , A) JP 63-261424 (JP, A) JP 7-129622 (JP, A) JP 7-146878 (JP, A) JP 10-260981 (JP, A) Masaki Yamashina, Junji Kojima, Interactive Sound Information Providing System, Information Processing Society of Japan, Research Report, Japan, Information Processing Society of Japan, January 26, 1990, Volume 90, No. 8, 1-7, Fujio Tsutsumi, Keyword Full-text search system that expresses search conditions by two-dimensional space arrangement, interactive system and software III Japan Society for Software Science and Technology WISS'95, Japan, Modern Science Co., Ltd., December 10, 1995, first edition, 221-229 page (58) investigated the field (Int.Cl. ^7, DB name) G06F 17/30 - 17/30 419 G06F 3/00 654 JICST file Le (JOIS)

Claims (15)

(57) [Claims] 1. A query output device for outputting a query given to a document search engine that obtains documents whose scores obtained by applying a query including a term to a document is a character string. A character string input acceptance unit that accepts the input of, a term extraction unit that morphologically analyzes the given character string and extracts the terms included in it, and when the character string input is accepted by the character string input acceptance unit An input control unit that gives the character string to the term extraction unit, a query generation unit that generates a query including the terms extracted by the term extraction unit, and a message display condition that meets the message display condition in advance. A message display unit that displays the associated message, and when the message is displayed by the message display unit, the character string input A correction input reception unit that receives a correction input for correcting the character string input by the reception unit; and, when the correction input is received by the correction input reception unit, the character string corrected by the correction input is used as the term extraction unit. a correction control section that gives, if the query output conditions are met, a query output unit for outputting a query generated by the query generation unit, and a term storage unit for storing a plurality of terms, by the query generation unit When the terms included in the generated query do not match any of the plurality of terms stored in the term storage unit, the message display condition is satisfied, and the query generation unit has a plurality of the extracted terms.
In this case, (a) terms of one general noun and terms of another general noun
And are concatenated (including the case where they are concatenated with the particle "no" in between)
Mu. ) Change the weight of these terms depending on
Generate a query, and (b) determine if one term contains another term or not.
Generating a query by changing the weights of these terms
Characteristic query output device. 2. The query generation unit, The term of the relevant general noun and the term of the other generic noun
When the and are connected, the weight of the term placed in front is
Make it bigger, If the one term includes the other term, the weight of the former
Is characterized by increasing the weight and decreasing the weight of the latter.
The query output device according to claim 1. 3. A parameter input acceptance unit for accepting an input of a generation parameter given to the query generation unit, wherein the query generated by the query generation unit is associated with each term in addition to the term, Including the weight referred to when obtaining the score, the query generation unit determines the magnitude difference of each weight of the terms extracted by the term extraction unit according to the generated parameter input of which is received by the parameter input reception unit. and, wherein the said term, the weights determined in association with this, to generate a query that contains a 請
The query output device according to claim 1 or 2 . 4. A parameter input acceptance unit for accepting an input of a generation parameter given to the query generation unit, wherein the query generated by the query generation unit is a term, and documents acquired by the document search engine are Including essential term information indicating whether or not to include the term, the query generation unit, each essential term information of the term extracted by the term extraction unit, the input is accepted by the parameter input acceptance unit The query output device according to claim 1 or 2, wherein the query output device determines a query based on the generated generation parameter, and generates a query including the term and the essential term information determined in association with the term. 5. The query output device according to claim 1, and a document in which a score obtained by applying the query output by the query output device to a document is included in a predetermined range. A document search system, comprising: a document search engine unit; and a document output unit that outputs a document acquired by the document search engine unit. 6. A query output method for outputting a query given to a document search engine that obtains documents whose scores obtained by applying a query including a term to a document is a character string. A character string input acceptance step of accepting the input of, a term extraction step of performing morphological analysis of the given character string and extracting the terms contained therein, and the character string input acceptance step of the character string input acceptance step In the case, an input control step of extracting a term from the character string in the term extraction step, a query generation step of generating a query including the term extracted in the term extraction step, and a message display condition are satisfied, A message display step of displaying a message previously associated with the message display condition; If a message is displayed in the message display step, a correction input reception step of receiving a correction input for correcting the character string input in the character string input reception step, and a correction input reception in the correction input reception step If the query output condition is satisfied, a modification control step of extracting a term from the character string modified by the modification input, and a query that outputs the query generated in the query generation step if the query output condition is satisfied and an output step, when the terms included in the query generated by the query generation step does not match any of the pre-stored plurality of terms, the message display condition is satisfied
In the query generation step, there are a plurality of extracted terms.
In some cases, (a) terms of one general noun and terms of another general noun
And are concatenated (including the case where they are concatenated with the particle "no" in between)
Mu. ) Change the weight of these terms depending on
Generate a query, and (b) determine if one term contains another term or not.
Generating a query by changing the weights of these terms
Characterized query output method. 7. In the query generation step, The term of the relevant general noun and the term of the other generic noun
When the and are connected, the weight of the term placed in front is
Make it bigger, If the one term includes the other term, the weight of the former
Is characterized by increasing the weight and decreasing the weight of the latter.
The query output method according to claim 6. 8. The method further comprises a parameter input accepting step of accepting an input of a generation parameter referred to in the query generating step, wherein the query generated in the query generating step corresponds to each term in addition to the term. The weight is referred to when obtaining the score, and the query generation step receives an input in the parameter input reception step regarding a difference in weight of each of the terms extracted in the term extraction step. The query output method according to claim 6 or 7, wherein a query including the term and the weight determined in association with the term is generated based on the generated generation parameter. 9. The method further comprises a parameter input accepting step for accepting an input of a generation parameter referred to in the query generating step, wherein the query generated in the query generating step is acquired by the document search engine in addition to the term. The document includes essential term information indicating whether or not the term is limited to those including the term, the query generation step, each essential term information of the terms extracted in the term extraction step, the parameter input acceptance step 8. The query according to claim 6 or 7, wherein the query is determined by the generation parameter received as input in step S6, and the term and the essential term information determined in association with the term are generated. output method. 10. A document search step of obtaining a document in which a score obtained by applying the query output by the query output method according to claim 6 to a document is within a predetermined range, And a document output step of outputting the document acquired in the document search step. 11. A computer, a character string input receiving unit for receiving a character string input, a term extracting unit for morphologically analyzing a given character string and extracting a term contained therein, the character string input receiving unit When a column input is accepted, an input control unit that gives the character string to the term extraction unit, a query generation unit that generates a query including the term extracted by the term extraction unit, and a message display condition is satisfied, A message display unit that displays a message previously associated with the message display condition, and a correction that receives a correction input that corrects the character string input by the character string input reception unit when the message is displayed by the message display unit The input acceptance unit, when the modification input is accepted by the modification input acceptance unit, More modified string, correction control unit that gives the term extractor, if the query output condition is satisfied, the query output unit for outputting a query generated by the query generation unit, Oyo
Beauty, to function as a term storage unit for storing a plurality of terms, wherein when a term included in the generated query by the query generation unit does not match any of the plurality of terms stored in said term storage unit, the message The display condition is
Satisfied, the query generator has multiple extracted terms
In this case, (a) terms of one general noun and terms of another general noun
And are concatenated (including the case where they are concatenated with the particle "no" in between)
Mu. ) Change the weight of these terms depending on
Generate a query, and (b) determine if one term contains another term or not.
A computer-readable recording medium, which records a program that functions to generate a query by changing the weight of these terms . 12. The method according to claim 12, The program is stored in the computer
Be careful The query generation unit, The term of the relevant general noun and the term of the other generic noun
When the and are connected, the weight of the term placed in front is
Make it bigger, If the one term includes the other term, the weight of the former
Function to increase the weight of the latter and decrease the weight of the latter.
Computer read according to claim 11, characterized in that
Possible recording medium. 13. The program further causes the computer to function as a parameter input reception unit that receives an input of a generation parameter given to the query generation unit, and the query generated by the query generation unit is a term in addition to the term. Corresponding to each of the terms, including the weight referred to when obtaining the score, the query generation unit, by the parameter input reception unit, the difference in weight of each of the terms extracted by the term extraction unit. A contract characterized in that the input is determined by the received generation parameter, and a query including the term and the weight determined in association with the term is generated.
The computer-readable recording medium according to claim 11 or 12 . 14. The program further causes the computer to function as a parameter input reception unit that receives an input of a generation parameter given to the query generation unit, and the query generated by the query generation unit is a term in addition to a term. The documents acquired by the search engine include essential term information indicating whether or not the documents include the term, and the query generation unit includes the essential term information of each of the terms extracted by the term extraction unit as the parameter. 13. The query, which is determined by a generation parameter whose input is accepted by the input accepting unit, and which includes the term and the essential term information determined in association with the term, is generated. Computer readable recording medium. 15. The document search engine unit, wherein the program obtains a document in which a score obtained by applying the query output by the query output unit to a document, is included in a predetermined range, by the computer. 15. The computer-readable recording medium according to claim 11, further functioning as a document output unit that outputs a document acquired by the document search engine unit.