JP2005018555A - Document data retrieval method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a document data retrieval method for sharply improving a correction rate and a cover rate. <P>SOLUTION: A computer for retrieving document data associated with a keyboard is made to execute a step 12 for making an alias dictionary preparing means prepare an alias dictionary by collecting an alias corresponding to a keyword from various public databases, a step 13 for making a document retrieval means retrieve and extract the document data where terms mentioned in the alias dictionary are used, a step 14 for making a simplified dictionary preparing means prepare a simplified character dictionary by extracting simplified character information from the document data, a step 15 for making a validity determination means check the terms used in the document data retrieved by the document retrieval means by using the alias dictionary and the simplified character dictionary, and determine the validity of the document data associated with the keyword and a step 16 for making an output means output the document data whose validity is determined by the validity determination means as a retrieval output result. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a document data search method and program for searching desired document data from a large number of document data using a keyword as a clue.
[0002]
[Prior art]
For example, when a biomedical researcher wants to collect information necessary for research activities, he / she can access online literature such as MEDLINE (bibliographic database storing biomedical articles) and obtain necessary information. Generally done. In this case, as a device for increasing the accuracy of the search result as much as possible, a technique of inputting a plurality of keywords is generally used.
[0003]
[Problems to be solved by the invention]
However, in recent years, the number of papers registered in a database such as MEDLINE has increased rapidly, and accordingly, it is difficult to obtain documents necessary for research accurately and efficiently. The main factor is the use of another name for gene or protein name. Even if researchers try to search for documents using gene or protein names as necessary keywords, many unrelated documents are included due to the ambiguity of names, and it is necessary for synonyms. There is a problem that documents are leaked from search results. In addition, since the input keywords have a great influence on the accuracy of search results, it is necessary to set appropriate keywords. However, advanced prior knowledge about search targets is required just by setting appropriate keywords. It may take time and effort. For this reason, when searching for documents required for research from biomedical academic papers, it is difficult to comprehensively and efficiently search only for documents containing the desired information, and the search results include many unrelated documents. It will be. Researchers need to select documents while confirming their contents, so a considerable amount of labor is required only by obtaining the documents. In addition, even if a document originally desired to be obtained is missing from the search, it is impossible to notice the fact.
The present invention is not intended to be applicable only to searches using gene names or protein names as keywords, and the search keywords may be anything other than gene names or protein names.
[0004]
An object of the present invention is to provide a document data search method and program capable of efficiently and comprehensively narrowing down required documents from a large number of documents using keywords.
[0005]
[Means for Solving the Problems]
In order to solve the above-mentioned problem, according to the present invention, there is provided a method for searching only document data related to a given keyword among a large number of document data using a computer, wherein the alias dictionary creating means Collecting aliases corresponding to the keywords from various public databases and creating an alias dictionary for the keywords; and searching and retrieving literature data in which the term search means uses the terms in the alias dictionary; Abbreviated dictionary creating means extracting abbreviation information from the document data extracted by the document retrieving means to create an abbreviation dictionary; and a validity determining means using the alias dictionary and the abbreviation dictionary for the document retrieval. Check whether the term used in the document data retrieved by the means is the document data related to the keyword And a document data search method comprising: a step of determining validity of the document; and a step of outputting the document data determined as having been validated by the validity determination unit as a search output result. Proposed.
[0006]
Further, an alias dictionary creating means for collecting an alias corresponding to the keyword from various public databases and creating an alias dictionary for the keyword to search for literature data related to the given keyword, Document retrieval means for retrieving and retrieving document data using the terminology used, Abbreviation dictionary creation means for creating abbreviation dictionary by extracting abbreviation information from document data retrieved by the document retrieval means, and the alias dictionary And abbreviation dictionary to check the terminology used in the document data searched by the document search unit and determine the validity of whether the document data is related to the keyword The document data determined to be valid by the validity determination means is output as a search output result. A literature data search program for functioning as a force means is proposed.
[0007]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an example of an embodiment of the present invention will be described in detail with reference to the drawings.
[0008]
FIG. 1 is a system configuration diagram of a document search apparatus according to the present invention, and FIG. 2 is a flowchart of a document search process executed by the document search apparatus.
[0009]
The document retrieval device 1 is configured by using a computer device, 2 is a central processing unit (CPU), 3 is a document database in which document data to be searched is stored, 4 is an external memory, 5 Is an internal memory in which a processing program for searching documents is stored, 6 is an input / output device, and these are connected to a bus 7. A keyboard 8, a display device 9, and a printing device 10 are connected to the input / output device 6. The document search apparatus 1 includes a communication control unit 7A so that various external databases can be accessed via the communication network N. The document search apparatus 1 is connected to the communication network N via a public line or the like. The document database 3A and the public database 3B connected to each other are accessed, and necessary documents can be fetched and stored in the external memory 4. In FIG. 1, only two databases, the document database 3A and the public database 3B, are illustrated, but the number of document databases may be any number.
[0010]
FIG. 2 is a diagram showing a processing flow of a document data search process performed by the processing program stored in the internal memory 5 being executed in the central processing unit 2.
[0011]
First, when a user inputs a keyword for document search from the keyboard 8 in step 11, in step 12, the communication control unit 7A accesses the document database 3A and the public database 3B via the communication network N, and so on. Find an alias for a given keyword used in a public database and create an alias dictionary. This alias dictionary is stored in the external memory 4. A file I for calling a given keyword is created from each alias, and the created file I is also stored in the external memory 4. This alias dictionary can create what alias is used for a given keyword from the public database by an appropriate means.
[0012]
In the next step 13, the communication control unit 7 </ b> A accesses the document database 3 via the communication network N, and, for example, from the document data stored in the corresponding biomedical document database, Search and download (collect) all document data that uses the term. The collected literature data is stored in the external memory 4. As described above, since a document search is performed using a large number of aliases for a given keyword as a search keyword, an improvement in the coverage ratio of the document search is expected. A search keyword group (which may be a single search keyword) that is the basis for collecting certain document data is created as a search keyword list (A) corresponding to the document data ID, and the list (A) is stored in an external memory. 4 is stored.
[0013]
In step 14, abbreviation information is extracted from each document data collected by the search process in step 13. The extracted abbreviation information is stored in the external memory 4 as an abbreviation dictionary corresponding to the document data ID. Here, the abbreviation information is information composed of a combination of an abbreviation appearing in the document data and a formal name (definition) corresponding to the abbreviation, and the abbreviation extraction process can be performed by using a known appropriate device or method. .
[0014]
In step 15, the validity of each document data searched in step 13 is determined using the alias dictionary, file I, list (A), and abbreviation dictionary stored in the external memory 4. That is, in the document set searched in step 13, there is a possibility that document data not related to a given keyword is collected due to the ambiguity of terms. Therefore, in order to improve the accuracy of the search, the validity of the searched document data is determined from the viewpoint of whether or not it can be said that the document data is related to the given keyword, and is not related to the given keyword as much as possible. Eliminate literature data. This validity determination is performed under the following conditions.
[0015]
When an abbreviation dictionary is created by the abbreviation extraction process from the document data to be determined,
1. The abbreviation portion of the abbreviation dictionary includes a search keyword that is the basis for collecting the document data, and includes at least one word that satisfies the criteria in the definition portion.
Or
A search keyword that is the basis for collecting the document data exists in the definition part of the abbreviation dictionary, and the abbreviation part includes at least one word that satisfies a certain criterion.
Here, a word satisfying a certain criterion is selected from all words in an alias dictionary for a given keyword.
2. The search keyword exists in the title or abstract (outline) part of the document data.
3. None of the words before and after the search keyword found in the title or the abstract part is a predetermined disallowed word.
Only when all the above three conditions are cleared, the document data is determined to be valid as the desired document data, and the process enters step 16.
[0016]
Search that is the basis for collecting document data in either abbreviations or definitions, even if an abbreviation dictionary is not created by the abbreviation extraction process from the document data to be determined, or even if an abbreviation dictionary is created If no keyword is included,
1. The title or abstract part of the document data contains at least one word that satisfies a certain standard.
However, when the search keyword is composed of only one word and satisfies a certain criterion, at least one word that satisfies a certain criterion other than the search keyword must be included.
Here, a word satisfying a certain criterion is selected from all words in an alias dictionary for a given keyword.
Or
The search keyword must consist of at least 6 characters.
However, the number of characters is not limited to this, and the setting can be changed as appropriate according to the search target.
2. The search keyword exists in the title or the abstract part.
3. None of the words before and after the search keyword found in the title or the abstract part is a predetermined disallowed word.
Only when all the above three conditions are cleared, the document data is determined to be valid as the desired document data, and the process enters step 16.
[0017]
As described above, the validity determination result for each document data is obtained, and only the document data determined to be valid in step 16 is output as the search result. Literature data that has not been determined to be valid is excluded from the search results. The output format of this search result can be freely combined with the ID (number), title, abstract, author, etc. of the document data according to the user's request. The output can be performed by displaying on the display device 9 and / or printing by the printing device 10. The search result can also be stored in the internal memory 5 so that it can be retrieved, displayed and printed at any time.
[0018]
Moreover, although the title or abstract part is used as an information object for document data discrimination, these are merely examples, and can be appropriately selected from all the information attached to the document data.
[0019]
3 to 5 show a more specific embodiment of the validity determination process executed in step 15. The validity determination processing program 20 first determines whether or not there is a validity between the search keyword that is the basis for collecting the document data and the given keyword, and the validity is determined from the search keyword list (A). A list (B) is created in which only the search keywords that have characteristics are left.
[0020]
In step 21, first, it is checked whether or not there is undetermined data in the document data to be validated. If it exists, the determination result in step 21 is YES and step 22 is entered. Thereafter, steps 22 to 35 are repeatedly executed for each document data until the document data becomes empty, and the validity of the document data is determined.
[0021]
In step 22, one document data is fetched from the external memory 4 into the internal memory 5. Subsequently, the abbreviation dictionary and list (A) corresponding to the document ID are fetched from the external memory 4 into the internal memory 5 respectively.
[0022]
In step 23, it is checked whether or not there is an undetermined search keyword in the list (A). If it exists, the determination result in step 23 is YES, and step 24 is entered. Thereafter, steps 24 to 32 are repeatedly executed for each search keyword until the list (A) becomes empty, and the validity of the search keyword is determined.
[0023]
In step 24, one search keyword is read from the list (A). Next, the file I corresponding to the search keyword is taken into the internal memory 5 from the external memory 4. Further, a given keyword corresponding to the search keyword is read from the file I, and an alias dictionary corresponding to the given keyword is taken into the internal memory 5 from the external memory 4. In the abbreviation dictionary corresponding to the imported document data, combination data of data (for example, abbreviation) within the parenthesis symbol and data immediately before the parenthesis symbol (for example, definition for the abbreviation) is recorded. ing. There may be a definition in parentheses and there may be an abbreviation just before the parenthesis. In any case, in the following description, data in parentheses is referred to as inner, and data immediately before the parentheses is referred to as outer.
There may be a plurality of given keywords corresponding to the search keyword in the file I. This is because, for the sake of ambiguity, the same search keyword is collected for a completely different given keyword. In this case, Steps 24 (capture the alias dictionary corresponding to the given keyword from the external memory 4 to the internal memory 5) to 32 are performed for all the given keywords, and then the process returns to Step 23.
[0024]
After fetching information necessary for discrimination in step 24, step 25 is entered. In step 25, it is determined whether or not there is an inner word of the abbreviation dictionary that matches the search keyword. If there is a matching inner, the determination result in step 25 is YES and step 26 is entered.
[0025]
In step 26, it is checked whether or not an alias matching the outer corresponding to the inner in the abbreviation dictionary exists in the alias dictionary corresponding to the given keyword, and the validity of the search keyword is determined. . If it is determined that there is validity, the determination result in step 26 is YES, the search keyword is stored in the list (B), and the process returns to step 23.
[0026]
If the determination result in step 26 is NO, step 27 is entered. In step 27, it is checked whether or not at least one word that matches an outer corresponding to the inner in the abbreviation dictionary is satisfied, and the validity of the search keyword is determined. If it is determined that there is validity, the determination result in step 27 is YES, the search keyword is stored in the list (B), and the process returns to step 23.
Here, a certain criterion is a word that is in the alias dictionary of the given keyword, is composed of four or more characters, and is other than a non-permitted word prepared in advance. . As described above, only words that satisfy all the criteria correspond to words that satisfy a certain criterion. However, the reference setting is not limited to this, and the setting can be appropriately changed according to the search target. If the decision result in the step 27 is NO, it is judged that the search keyword is unrelated to the given keyword, and the process returns to the step 23 as it is.
[0027]
If the determination result in step 25 is NO, step 28 is entered. In step 28, it is determined whether or not there is a match with the search keyword in the outer part of the abbreviation dictionary. If there is a matching outer, the determination result in step 28 is YES and step 29 is entered.
[0028]
In step 29, it is checked whether or not an alias matching the inner corresponding to the outer of the abbreviation dictionary exists in the alias dictionary corresponding to the given keyword, and the validity of the search keyword is determined. . If it is determined that there is validity, the determination result in step 29 is YES, the search keyword is stored in the list (B), and the process returns to step 23.
[0029]
If the determination result in step 29 is NO, step 30 is entered. In step 30, it is checked whether or not at least one word that matches the inner corresponding to the outer of the abbreviation dictionary and satisfies a certain criterion is included, and the validity of the search keyword is determined. If it is determined that there is validity, the determination result in step 30 is YES, the search keyword is stored in the list (B), and the process returns to step 23.
Here, the certain standard is the same as the standard used in step 27. If the determination result in step 30 is NO, it is determined that the search keyword is unrelated to the given keyword, and the process returns to step 23 as it is.
[0030]
If the determination result in step 28 is NO, that is, if it is determined that the keyword is not included in either the inner or outer by referring to the abbreviation dictionary, or there is no abbreviation dictionary corresponding to the document data. If so, step 31 is entered. In step 31, it is checked whether the title or abstract part of the document data contains at least one word that satisfies a certain criterion, and the validity of the search keyword is determined. If it is determined that there is validity, the determination result in step 31 is YES, the search keyword is stored in the list (B), and the process returns to step 23.
Here, the certain standard is the same as the standard used in step 27.
[0031]
If the determination result in step 31 is NO, the process enters step 32. In step 32, it is checked whether or not the search keyword is composed of 6 characters or more, and the validity of the search keyword is determined. If it is determined that there is validity, the determination result in step 32 is YES, the search keyword is stored in the list (B), and the process returns to step 23. If the determination result in step 32 is NO, it is determined that the search keyword is unrelated to the given keyword, and the process returns to step 23 as it is.
[0032]
If the decision result in the step 23 is NO, that is, if there is no undetermined search keyword in the list (A), the process enters the step 33. In step 33, it is checked whether or not there is an undetermined search keyword in the list (B). If it exists, the determination result in step 33 is YES, and step 34 is entered. Thereafter, steps 34 to 35 are repeatedly executed for each search keyword until the list (B) becomes empty, and the validity of the search keyword is determined.
[0033]
In step 34, one search keyword is read from the list (B). Next, it is determined whether or not the search keyword exists in the title or abstract of the document data. If the search keyword exists in the document data, the determination result in step 34 is YES and the process enters step 35. If the determination result in step 34 is NO, the process returns to step 33.
[0034]
In step 35, it is checked whether or not the words before and after the search keyword found in the title or abstract of the document data are predetermined disallowed words, and neither of the words before or after corresponds to the disallowed word. If YES in step 35, the determination result in step 35 is YES, the combination of the document ID and the given keyword is saved in the list II as a search result, and the process returns to step 33.
However, the words before and after the search keyword are limited to the same sentence, and when the search keyword is used at the beginning or end of a sentence, only the word immediately after or immediately before is targeted.
[0035]
If the decision result in the step 35 is NO, that is, if any of the search keywords before or after the search keyword corresponds to an unauthorized word, the process returns to the step 33 as it is.
[0036]
If the determination result in step 33 is NO, that is, if there is no undetermined search keyword in the list (B), the process returns to step 21.
[0037]
In this way, when the search processing for all prepared processing target sentence data is completed, the determination result in step 21 is NO, and the processing of this program is completed. The search result is presented to the user after extracting information necessary for output from the list II.
[0038]
【The invention's effect】
According to the present invention, it is possible to dramatically improve the search accuracy, that is, the accuracy rate and the coverage rate, and it is possible to greatly reduce the researcher's information collection time and labor.
[Brief description of the drawings]
FIG. 1 is a configuration diagram showing an example of an embodiment of the present invention.
FIG. 2 is a flowchart showing a processing program executed in the search processing system.
FIG. 3 is a diagram showing a part of a detailed flowchart of validity determination processing in FIG. 2;
FIG. 4 is a diagram showing a part of a detailed flowchart of validity determination processing in FIG. 2;
FIG. 5 is a view showing a part of a detailed flowchart of validity determination processing in FIG. 2;
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Literature retrieval device 2 Central processing unit 3 Literature database 4 External memory 5 Internal memory 6 Input / output device 7 Bus 8 Keyboard 9 Display device 10 Printing device

Claims (2)

A method for searching only literature data related to a given keyword among a large number of literature data using a computer,
An alias dictionary creating means collecting aliases corresponding to the keywords from various public databases and creating an alias dictionary for the keywords;
A step in which a document retrieval means retrieves and retrieves document data in which terms in the alias dictionary are used;
Abbreviated dictionary creating means for extracting abbreviation information from the document data extracted by the document search means to create an abbreviation dictionary;
The validity determination means checks the terms used in the document data searched by the document search means using the alias dictionary and the abbreviation dictionary, and whether or not the document data is related to the keyword. Determining gender,
A document data search method comprising: a step of outputting document data determined to be valid by the validity determination unit as a search output result. A computer to search for literature data related to a given keyword,
Alias dictionary creating means for collecting aliases corresponding to the keywords from various public databases and creating an alias dictionary for the keywords;
Literature retrieval means for retrieving and retrieving literature data in which terms in the alias dictionary are used,
Abbreviation dictionary creation means for creating abbreviation dictionary by extracting abbreviation information from the literature data extracted by the literature search means;
Validity for determining whether or not the document data is related to the keyword by checking the terms used in the document data searched by the document search means using the alias dictionary and the abbreviation dictionary Sex discrimination means,
A literature data search program for functioning as output means for outputting literature data determined to be valid by the validity judgment means as a search output result.

Images