JP2013041385A - Document retrieval method, document retrieval device, and document retrieval program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a document retrieval method which efficiently retrieves a patent document by reducing retrieval omission.SOLUTION: A document retrieval device performs the steps of: on the basis of inputted retrieval keywords, retrieving documents including the retrieval keywords from database in which retrieval object documents are accumulated; extracting first keywords from the first retrieved result; extracting a classification codes included in a group of documents in the retrieved result; tabulating the extracted classification codes; retrieving documents including the classification codes from the database on the basis of the tabulated classification codes; extracting second keywords included in a group of documents in the second retrieval result; comparing the first keywords with the second keywords to extract keywords that are not included in the first keywords; and displaying the first retrieval result and the keywords in the keyword comparison result.

Description

  The present invention relates to a document search method, a document search apparatus, and a document search program for searching for a document.

  In order to file a patent application, it is essential to search for prior art. Usually, the inventor uses a patent search system represented by IPDL to search the prior art. At this time, the inventor generally performs a search using a keyword reminiscent of his invention. The inventor anticipates a search keyword that hits the desired patent gazette, repeats the search while devising the input method, and finds useful information.

  By the way, in a general search system, one or a plurality of search keywords are input, and those keywords are all included in the document (logical product or AND search), some are included (logical sum or OR search), Or it is possible to include logical operations such as nothing (negation or NOT search). As a result, the target of the document group expected by the user can be brought closer. In addition, many search systems that support a combination of logical operations, such as those containing either keyword A or B and also containing C, are supported.

  However, several problems arise when an actual search is performed using such a search technique. For example, let us consider a case where a search is made by specifying keywords “camera” and “zoom lens” in order to search for prior art regarding a zoom lens of a camera. At this time, if the term “high-power lens” is used instead of “zoom lens” in the text of the publication, the search will not be hit. That is, if such a synonym or a keyword including the same concept is not conceived at the time of retrieval, information may be missed. On the other hand, suppose that the keywords “camera”, “zoom lens”, and “high-power lens” are conceived and the range of the search target is expanded to include any of them. However, as a result, if thousands or tens of thousands of search results are presented, there may be many publications that the inventor does not want, and it is practically impossible to read all of them. Alternatively, there is a possibility that a document to be investigated as a known example does not use any keyword. In other words, even if the search results are not missed (or few), if there are too many documents that hit the search, it is practical to limit the number of documents to a suitable number by narrowing down to some extent. On the other hand, if there are no documents that hit the search, or if there are too few documents, it is necessary to propose a keyword that is associated with the search keyword input by the user to expand the search possibility.

  As a narrowing down method, there is a search method using a classification code. For example, in the case of patent information, hierarchical patent codes such as IPC (International Patent Classification), Japanese FI (File Index) and F (File Forming) terms, and US Patent Classification are included in each patent document. Has been granted. In particular, in FI and F-term, the classification code values are subdivided into 100,000 items or more. Therefore, when the user can know the value of the classification code suitable for the purpose and purpose of the search, the search can be performed using the value of the classification code as a key to perform a high-precision search with less search noise and search leakage. It can be realized. However, since the classification is hierarchized and subdivided, it is difficult for the inventor to select a classification code that is unfamiliar and find a classification code suitable for the search purpose and intention.

  In Patent Document 1, a keyword dictionary is prepared, a keyword related to the input keyword is derived, and the inventor's search is assisted.

  In Patent Document 2, a classification code is extracted from a search result, a classification code to be used for narrowing down is derived, and the inventor's search is protected.

JP 2010-003015 A JP 2008-165401 A

As described above, it is possible to deal with search omissions and narrowing down by deriving related keywords and classification codes.
(Problem 1) The keyword dictionary is prepared by the system provider and is not always the latest information. Therefore, even if a keyword that should be newly related appears, it is reflected in the dictionary. However, the keyword may not be derived.
(Problem 2) The classification code is still unfamiliar to the inventor, and even if an effective classification code is found for the search result, there is still a threshold to use it for the next search. It does not always contribute greatly to the efficiency of the inventors' search.

  The present invention has been made in order to solve the above-described problems. A document search method capable of efficiently performing a search by reducing search omissions at the time of patent document search, a document search apparatus for performing the search, and An object is to provide a literature search program.

  One means for solving the above problem will be described. The first aspect of the present invention is to search for a document including a corresponding search keyword from a database in which documents to be searched are accumulated based on the input search keyword, and store the search result in a storage device. It is the literature search method in the literature search device which performs search processing. Then, the document search device extracts a first keyword from the result of the first search process, and extracts a classification code included in a document group as a result of the first search process, and an extraction Based on the classified and aggregated processing for storing the totaled classification codes and storing the totaled results in the storage device, and on the basis of the totaled classification codes, the documents including the corresponding classification codes are searched from the database. A second search process stored in the storage device; a keyword extraction process for extracting a second keyword included in a document group of search results searched by the second search process; the first keyword; A keyword comparison process for comparing a second keyword and extracting a keyword that does not exist in the first keyword, and a search result and a key obtained by the first search process Which comprises carrying out the search result display process for displaying the keywords over de comparison result.

  According to the present invention, it is possible to reduce search omission of patent search and to perform search efficiently.

It is a figure which shows an example of the whole structure of a literature search system. 2 is a diagram illustrating an example of a hardware configuration of a document search apparatus 102. FIG. 5 is a flowchart illustrating an example of an outline of operation of a document search apparatus 102. It is a figure which shows an example of the record of a gazette table. It is a figure which shows the example of a search result group. It is a figure which shows the example of a search keyword group. It is a figure which shows the example of a classification code group. It is a figure which shows the example of a re-search result group. It is a figure which shows the example of the search keyword group by re-search. It is a figure which shows the example of the group of a related keyword. It is a figure which shows the example of the process of search result group A and keyword group alpha production | generation. It is a figure which shows the example of a display result.

  Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

  First, the overall configuration of the document search system according to the present embodiment will be described.

  FIG. 1 is a configuration diagram schematically showing the overall configuration of a document search system constructed as an embodiment of the present invention. This system includes a terminal 101 operated by the inventor, a document search device (server) 102 on which the system operates, and a network 103 connecting the terminal 101 and the document search device 102.

  FIG. 2 is a diagram illustrating an example of a hardware configuration of the document search apparatus 102 constructed as an embodiment of the present invention. The document search apparatus 102 can be realized by a general computer (electronic computer), and includes a communication apparatus 200 that transmits and receives information to and from the terminal 101 via the network 103, and an input apparatus for the inventors to input search keywords. 201, a display device 202 that displays search results, a processing device 102, a storage device 203, and a literature database (hereinafter, “database” is referred to as “DB”) 204. Each device is connected by a bus (BUS) or the like. The processing device 102 includes a search execution processing unit 205, a search result display processing unit 206, a classification extraction processing unit 207, a classification totaling processing unit 208, a re-search processing unit 209, a keyword extraction processing unit 210, and a keyword comparison processing unit 211. Here, the input device 201 and the display device 202 may not be included in the document search device 102. Further, the storage device 203 and the document DB 204 do not necessarily have to be inside the document search device 102, and may be arranged in another device. The processing device 102 can be realized by a CPU or the like. The processing device 102 loads various programs from the storage device 203 to a memory (not shown) and executes predetermined programs to realize each processing unit of the above. That is, the processing units 205 to 212 are realized as a processing process of the processing device 102. Predetermined programs for realizing each processing unit may be stored in the storage device 203 in advance, or stored in a portable storage medium that can be used by a computer, and read through the reader as needed. Or may be downloaded as necessary from the network 103, which is a communication medium that can be used by a computer, or another device connected to the communication device 200 using a carrier wave propagating on the network 103, and stored. It may be stored in the device 203.

  The storage device 203 stores a keyword group α501, a search result group A601, a classification code 701, a search result group B801, a keyword group β901, and a related keyword 1001, which will be described later. The document DB 204 stores a bulletin table that stores records having information such as the body text 402, IPC 403, FI 404, and F term 405 for each bulletin.

  The search execution processing unit 205 executes search processing based on a search request from the terminal 101 or the input device 201.

  The search result display processing unit 206 transmits the search result screen to the search request source terminal 101 or displays it on the display device 202.

  The classification extraction processing unit 207 extracts a classification code included in the document group of the search result.

  The classification totalization processing unit 208 totalizes the extracted classification codes.

  The re-search processing unit 209 searches the document DB 204 for documents including the corresponding classification code based on the collected classification codes.

  The keyword extraction processing unit 210 extracts keywords included in a document group of search results found by re-searching.

  The keyword totalization processing unit 211 totals the keywords extracted by the keyword extraction processing unit 210.

  The keyword comparison processing unit 212 compares the keywords aggregated by the keyword aggregation processing unit 211 with a keyword group α501 described later.

  Next, the operation of the document search system according to this embodiment will be described.

  FIG. 3 is a flowchart showing an example of the outline of the operation of the document search system according to the present embodiment.

  In steps 301 and 302, the search execution process 205 receives a search request including a search keyword input by the inventors from the input device 201 or the terminal 101, and the search process is executed based on the received search keyword.

  A series of processing in step 303 is shown in FIG. In the search execution processing 205, a full-text search is performed on the text 402 of the gazette table in FIG. 4 by the input search keyword 1101, and the hit records are collected and configured by the publication number 401 in the gazette table in FIG. 4. A search result group A 501 in FIG. 5 is generated on the storage device 201. Further, referring to the publication table of FIG. 4 from the public number of the search result group A 501, the record is identified from the publication number 401, the keyword in the text 402 is acquired, and the keyword group α 601 of FIG. 6 is stored in the storage device 201. Generated. Acquisition of keywords is performed by morphological analysis. The publication table in FIG. 4 is a table prepared in advance by the system, and information on documents is accumulated.

  In steps 304 and 305, the classification extraction process 207 refers to the publication table of FIG. 4 for each document in the search result group A501, identifies the record from the publication number 401, and extracts the IPC 403 as classification data ( 7, the totalization is performed, and the patent classification group 701 of Fig. 7 is generated on the storage device 201. At this time, the FI 404 or the F term 405 can be used as the classification data. It is also possible to generate a group by narrowing down only the top classifications with a large number of granted publications as a result of aggregation.

  In steps 306 and 307, a search is executed in the re-search process 209 using the classification code in the patent classification group 701, the publication number 401 of the hit document is extracted, and the search result group B801 in FIG. Generated on top. Here, when a number such as an application number or a registration number is stored in the document DB 204 as a number 401 other than the publication number, it may be extracted.

  In steps 308 and 309, in the keyword extraction process 210, for each document in the search result group B 801, the record is identified from the publication number 401 with reference to the publication number 401, and the keyword is obtained from the body text 402 through morphological analysis. 9 is extracted, and totalization is performed in the keyword totalization processing 211, and the keyword group β901 in FIG. 9 is generated on the storage device 201. At this time, it is also possible to generate a group by narrowing down to only upper keywords having a high appearance frequency as a result of aggregation.

  In steps 310 and 311, the keyword group α 501 and the keyword group β 901 are compared by the keyword comparison process 212, and keywords that exist in the keyword group β 901 and do not exist in the keyword group α 501 are extracted, and the related keyword group 1001 in FIG. Is generated on the storage device 201.

  In step 312, the search result display process 206 displays the search result group A 601 and the related keyword group 1001 on the display device 202. A display example is shown in FIG.

  The embodiment of the present invention has been described above.

  According to the above embodiment, search omission at the time of patent document search can be reduced and search can be performed efficiently. Further, since the search result is displayed and related keywords are displayed, it is possible to efficiently and directly understand what keywords should be added to the narrowing down. Moreover, since it also becomes a material for determining what keyword should be used for the search again, it is possible to efficiently select keywords to be used for the next search.

  As mentioned above, although the Example of this invention was described, this invention is not limited to this, A various change is possible in the range which does not deviate from the summary.

DESCRIPTION OF SYMBOLS 101 ... Terminal, 102 ... Network, 103 ... Document search device (server), 201 ... Input device, 202 ... Display device, 203 ... Memory, 204 ... Literature DB.

Claims (3)

Based on the input search keyword, search the document containing the search keyword from the database that stores the documents to be searched, and implement the first search process to store the search result in the storage device A document search method in a document search device that performs:
Extracting a first keyword from the result of the first search process;
A classification extraction process for extracting a classification code included in a document group as a result of the first search process;
A classification totaling process for totalizing the extracted classification codes and storing the totalization results in the storage device;
Based on the aggregated classification codes, search for documents including the corresponding classification codes from the database, and storing the search results in the storage device,
A keyword extraction process for extracting a second keyword included in a document group of search results searched by the second search process;
A keyword comparison process for comparing the first keyword with the second keyword and extracting a keyword that does not exist in the first keyword;
A search result display process for displaying the search result and the keyword of the keyword comparison result obtained in the first search process;
A document retrieval method characterized by that. Based on the input search keyword, a first search means for searching for a document including the corresponding search keyword from a database in which documents to be searched are stored and storing the search result in a storage device is provided. A document retrieval device,
Means for extracting a first keyword from a search result by the first search means;
Classification extraction means for extracting a classification code included in the document group of the search results;
A classification totaling unit that totalizes the extracted classification codes and stores the totalization results in the storage device;
Based on the totaled classification code, a second search means for searching a document including the corresponding classification code from the database, and storing the search result in the storage device;
Keyword extracting means for extracting a second keyword contained in a document group of search results searched by the second search means;
A keyword comparing means for comparing the first keyword with the second keyword and extracting a keyword that does not exist in the first keyword;
Search result display means for displaying the search results obtained by the first search means and the keywords of the keyword comparison results;
A document retrieval apparatus characterized by that. A first search in which a computer is searched for a document including the corresponding search keyword from a database in which documents to be searched are stored based on the input search keyword, and the search result is stored in a storage device. A literature search program for executing processing,
Extracting a first keyword from the result of the first search process;
A classification extraction process for extracting a classification code included in a document group as a result of the first search process;
A classification totaling process for totalizing the extracted classification codes and storing the totalization results in the storage device;
Based on the aggregated classification codes, search for documents including the corresponding classification codes from the database, and storing the search results in the storage device,
A keyword extraction process that extracts and aggregates second keywords included in the document group of the search results searched by the second search process, and stores the aggregation results in the storage device;
A keyword comparison process for comparing the first keyword with the second keyword and extracting a keyword that does not exist in the first keyword;
A search result display process for displaying a search result obtained in the first search process and a keyword of a keyword comparison result is executed.
A document search program characterized by that.

Images