JP6253352B2 - Document analysis support system - Google Patents

Description

  The present invention relates to a data processing technique, and more particularly to a technique for supporting analysis of document data.

  When a company decides the direction of R & D, it is important to clarify the position of its core technology and to grasp the technological development trends of competitors. It is useful to conduct an analysis of Conventionally, a technology is known in which patent documents to be analyzed are acquired, analysis processing such as morphological analysis and dependency analysis is performed on the summary and claims of each patent document, and the analysis results are statistically analyzed. (For example, Patent Document 1).

JP2011-257817A

  Analysis processing such as morphological analysis and dependency analysis is a processing with a relatively high load. For this reason, when there are a large number of patent documents to be analyzed, this analysis process may take a long time. This hinders the provision of the above-described technology as an ASP-type service.

  The present invention has been made in view of these problems, and an object thereof is to provide a document analysis support system that can analyze patent documents in a relatively short time.

  In order to solve the above problems, a document analysis support system according to an aspect of the present invention includes a morpheme information holding unit that holds morpheme information obtained by dividing each of a plurality of document data held in a database into morphemes, and the database holds A list acquisition unit that acquires a list of document data to be analyzed from the document data to be analyzed, and an analysis information extraction unit that extracts morpheme information of the document data included in the list from the morpheme information holding unit.

  It should be noted that any combination of the above-described constituent elements, and the expression of the present invention converted between a method, a system, a program, a recording medium storing the program, and the like are also effective as an aspect of the present invention.

  According to the present invention, patent documents can be analyzed in a relatively short time.

It is a figure which shows the structure of the document analysis assistance system of embodiment. It is a block diagram which shows the function structure of the document analysis assistance apparatus of FIG. It is a data structure figure which shows morpheme information. It is a data structure figure which shows dependency information. It is a flowchart which shows a series of processes which concern on the analysis process in the document analysis assistance apparatus of FIG. It is a flowchart which shows a series of processes which concern on the extraction and totalization process in the document analysis assistance apparatus of FIG.

The outline of the document analysis support system according to the present embodiment is as follows.
The document analysis support system according to the present embodiment holds in advance the results of morphological analysis of all patent documents held in the patent database. In addition, the result of the dependency analysis using the analysis result is stored in advance. The document analysis support system searches the patent database under a predetermined search condition, and acquires a list of patent documents to be analyzed. For example, when a patent application of “Applicant A” is to be analyzed, “Applicant A” is set in “Applicant / right holder” of the search condition, and the patent database is searched. The document analysis support system extracts morpheme information and dependency information on patent documents included in the list from morpheme information and dependency information held in advance. Then, they are aggregated and analyzed. That is, in the document analysis support system according to the present embodiment, since the patent document to be analyzed is previously subjected to morphological analysis and dependency analysis, it is not necessary to perform analysis when analyzing. When analyzing, it is sufficient to acquire the results of the analysis and aggregate them. Therefore, it becomes possible to analyze patent documents in a relatively short time.

  FIG. 1 shows a configuration of a document analysis support system 10 according to the embodiment. The document analysis support system 10 includes a document analysis support device 100, a patent database 200, and a user terminal 300. Each of these devices is connected via a known communication network such as LAN, WAN, or the Internet.

  The document analysis support apparatus 100 is an apparatus for supporting the analysis of patent documents held in the patent database 200. A detailed functional configuration of the document analysis support apparatus 100 will be described later with reference to FIG. The user terminal 300 is an information processing terminal operated by a user. The user terminal 300 is a general PC (Personal Computer) terminal in which a web browser is installed. The user terminal 300 may be a tablet terminal or a smartphone.

  The patent database 200 holds patent documents that have been published. The patent document referred to here includes accompanying bibliographic information such as progress information. The patent database 200 holds patent documents in a text format. The patent database 200 may be an existing database such as an NRI Cyber Patent Desk (registered trademark) or an electronic patent library provided by an industrial property information / training hall.

  FIG. 2 is a block diagram showing a functional configuration of the document analysis support apparatus 100 of FIG. Each of these blocks can be realized in hardware by an element such as a CPU of a computer or a mechanical device, and is realized in software by a computer program or the like, but here, functions realized by their cooperation. Draw a block. Therefore, those skilled in the art will understand that these functional blocks can be realized in various forms by a combination of hardware and software.

  The document analysis support apparatus 100 includes an analysis unit 120, a patent document search unit 130, a patent document list acquisition unit 140, an analysis information extraction unit 150, a totaling unit 160, a display control unit 170, and a morpheme information holding unit 181. A dependency information holding unit 182, an extracted morpheme data holding unit 183, an extracted dependency data holding unit 184, a dictionary data holding unit 185, a morpheme total data holding unit 186, and a dependency total data holding unit 187. .

  The analysis unit 120 analyzes the patent document. The analysis unit 120 includes a morpheme analysis unit 121 and a dependency analysis unit 122. The morpheme analysis unit 121 performs morpheme analysis of all patent documents held in the patent database 200. Here, the morpheme analysis means that a sentence is divided into a character string (morpheme) of the smallest unit having meaning, and the divided character string is classified into parts of speech. For example, when “my name is Suzuki” is broken down into morphemes, “I (pronoun)” “no (particle)” “name (general noun)” “ha (counselor)” “Suzuki (proprietary noun)” “ Is (auxiliary verb) ".

  The morpheme analysis unit 121 performs morpheme analysis with reference to the dictionary data held in the dictionary data holding unit 185. As a result, the patent document can be more accurately divided into morphemes. This will be described later. In the present embodiment, the morpheme analysis unit 121 performs morpheme analysis on the subject of the summary of each patent document and claim 1 of the claims. In addition to or in addition to these, each means of the summary solution, other claims of the claims, and the specification may be subjected to morphological analysis. The morpheme information obtained as a result of the morpheme analysis is held in the morpheme information holding unit 181. The morpheme information will be described later with reference to FIG.

  In addition, when the patent document held in the patent database 200 is updated, the morpheme analysis unit 121 performs morphological analysis of the updated patent document. Specifically, when a patent document is updated by correcting the specification, etc., when a patent application is newly published and the patent document is added to the patent database 200, the morphological analysis unit 121 updates / adds them. The morpheme analysis of the published patent document is executed, and the morpheme information held in the morpheme information holding unit 181 is updated. In general, the patent database 200 is regularly updated at a frequency of once every two weeks. Scheduling may be performed so that the morphological analysis unit 121 executes the morphological analysis of the patent document updated / added after the periodic update. Of course, the morphological analysis may be executed after receiving the instruction.

  In addition, the morphological analysis unit 121 performs morphological analysis even when the dictionary data is updated. In this case, since the analysis results of all patent documents may be affected, the morpheme analysis unit 121 executes morpheme analysis of all patent documents again, and the morpheme information held in the morpheme information holding unit 181. Update.

  The dependency analysis unit 122 performs dependency analysis using the analysis result in the morpheme analysis unit 126, that is, morpheme information. Specifically, the dependency between morphemes is determined. The dependency information obtained as a result of the dependency analysis is held in the dependency information holding unit 182. The dependency information will be described later with reference to FIG. Similar to the morphological analysis unit 121, the dependency analysis unit 122 executes dependency analysis of the updated patent document when the patent document stored in the patent database 200 is updated, and stores the dependency in the dependency information storage unit 182. Update the dependency information. When the dictionary data is updated, dependency analysis of all patent documents is executed again, and dependency information is updated. Note that the dependency analysis unit 122 may be configured to start the dependency analysis upon completion of the morphological analysis by the morpheme analysis unit 121.

  The patent document search unit 130 searches for patent documents to be analyzed. The patent document search unit 130 includes a search condition acquisition unit 131, a search expression generation unit 132, and a search execution unit 133. The search condition acquisition unit 131 receives from the user terminal 300 input of search conditions for searching for patent documents to be analyzed. Applicant / right holder name, applicant identification number, inventor name, publication date, filing date, filing date, keywords for searching for forms, claims and abstracts, etc. Search conditions can be entered arbitrarily. For example, to analyze a patent application in the technical field of “satellite positioning” of “applicant A”, the name of the applicant / right holder is “applicant A”, the form for carrying out the invention, the scope of claims, Enter "satellite positioning" as a search keyword such as summary.

  The search expression generation unit 132 generates the search expression 21 based on the search conditions received by the search condition acquisition unit 131. The search execution unit 133 searches the patent database 200 based on the search formula 21 generated by the search formula generation unit 132.

  The patent document list acquisition unit 140 obtains a list of patent documents obtained as a result of the search by the search execution unit 133 or a list of patent documents specified from the user terminal 300 (hereinafter collectively referred to as the patent document list 22). It is acquired in the form of a document ID such as

  The analysis information extraction unit 150 extracts the analysis result of the patent document. The analysis information extraction unit 150 includes a morpheme information extraction unit 151 and a dependency information extraction unit 152. The morpheme information extraction unit 151 extracts the morpheme information of the patent document specified by the document ID included in the patent document list 22 from the morpheme information holding unit 181. The morpheme information extracting unit 151 records the extracted morpheme information in the extracted morpheme data holding unit 183.

  The dependency information extraction unit 152 extracts the dependency information of the patent document specified by the document ID 24 included in the patent document list 22 from the dependency information holding unit 182. The dependency information extracting unit 152 records the extracted dependency information in the extracted dependency data holding unit 184.

  The totaling unit 160 totalizes the extracted analysis results. The totaling unit 160 includes a morpheme totaling unit 161 and a dependency totaling unit 162. The morpheme totaling unit 161 totals the appearance frequency of each morpheme included in the morpheme information held in the extracted morpheme data holding unit 183 while grouping them with arbitrary items. The user can set which items are grouped and aggregated. For example, the morpheme totaling unit 161 may total the appearance frequency of each morpheme while grouping by applicant. In this case, it is possible to grasp technical terms that appear frequently in each of the applicants' rejected documents. In addition, technical terms that appear only in the patent documents of a specific applicant can be grasped. And, from these, it is possible to grasp the technology that each applicant is placing importance on.

  Further, for example, the morpheme totaling unit 161 may total the appearance frequency of each morpheme while grouping by ages and applicants. In this case, technical terms having a high appearance frequency can be grasped for each age and for each applicant, and changes in the technical trends of each applicant can be grasped. The morpheme totaling unit 161 records the totalization result in the morpheme total data holding unit 186.

  The dependency totaling unit 162 totals the appearance frequencies of each dependency relationship included in the morpheme information held in the extracted dependency data holding unit 184 while grouping them with arbitrary items. The user can set which items are grouped and aggregated. For example, the dependency totaling unit 162 may total the appearance frequencies of morphemes having a dependency relationship with a specific morpheme (for example, “issue”) while grouping by applicant.

  Further, for example, the dependency totaling unit 162 may total the appearance frequencies while grouping morphemes that are in a dependency relationship with a specific morpheme by synonyms. For example, it has a dependency relationship with the phrase “issue” and is synonymous with “usability”, “ease of viewing”, “confidentiality”, “improving accuracy”, “smaller and lighter”, and “low power consumption”. You may total the appearance frequency of a morpheme for every five character strings. In this case, it is possible to grasp which of “usability”, “easy to read”, “confidentiality”, “accuracy improvement”, “small and light weight”, and “low power consumption” is an issue. The dependency totaling unit 162 records the totaling result in the dependency totaling data holding unit 187.

  The display control unit 170 causes the user terminal 300 to display the results obtained by the aggregation unit 160. In addition, the display control unit 170 performs principal component analysis on the morpheme information held in the morpheme information extraction unit 151 and the dependency information held in the dependency information extraction unit 152, and the morpheme information and dependency information May be visualized by arranging them on a two-dimensional map.

  The morpheme information holding unit 181 holds morpheme information. In particular, the morpheme information holding unit 181 holds morpheme information for all patent documents held in the patent database 200. FIG. 3 is a data structure diagram showing morpheme information. The document ID 24 indicates an ID that uniquely identifies a patent document. Item 28 indicates an item including each morpheme in the patent document. The sentence number 30 indicates a number that uniquely identifies a sentence included in each item within the item. The morpheme ID 32 indicates an ID that uniquely identifies a morpheme included in each sentence within the sentence. The morpheme 34 indicates a morpheme included in each sentence. The part of speech 36 indicates the part of speech of the morpheme. For example, the first sentence of the summary of a patent document with a document ID “JP 2003-0001” includes a morpheme “lever (noun)”.

  The dependency information holding unit 182 holds dependency information. In particular, the dependency information holding unit 182 holds dependency information for all patent documents held in the patent database 200. FIG. 4 is a data structure diagram showing dependency information. A morpheme ID (relation source) 40 and a morpheme (relation source) 42 indicate a morpheme ID and a morpheme, respectively. The morpheme ID (relation destination) 44 and the morpheme (relation destination) 46 indicate the morpheme ID and morpheme of the relationship destination, respectively. For example, it shows that “tape (noun)” and “print (noun)” in the first sentence of the patent document invention whose document ID is “JP 2003-0001” are in a dependency relationship.

  The extracted morpheme data holding unit 183 holds the morpheme information of the patent document to be analyzed, that is, the morpheme information extracted from the morpheme information holding unit 181 by the morpheme information extracting unit 151. The data structure of the information held by the extracted morpheme data holding unit 183 is the same as that shown in FIG.

  The extracted dependency data holding unit 184 holds dependency information of the patent document to be analyzed, that is, dependency information extracted from the dependency information holding unit 182 by the dependency information extracting unit 152. The data structure of the information held by the extraction dependency data holding unit 184 is the same as that shown in FIG.

  The dictionary data holding unit 185 holds dictionary data having technical terms and technical terms. By referring to this dictionary data, the morpheme analyzer 121 can accurately divide the data into morpheme units. For example, if the dictionary data has the term “gas-liquid interface”, which is a technical term that means the interface between gas and liquid, if this term is included in the patent literature, “gas-liquid” And “interface” can be treated as one word “gas-liquid interface”.

  The morpheme total data holding unit 186 holds the data totaled by the morpheme total unit 161. The dependency total data holding unit 187 holds data totaled by the dependency totaling unit 162.

The operation of the document analysis support apparatus 100 having the above configuration will be described.
FIG. 5 is a flowchart showing a series of processes related to the analysis process in the document analysis support apparatus 100. The morphological analysis unit 121 has not performed morphological analysis on the patent document held in the patent database 200 (Y in S10), or the dictionary data held in the dictionary data holding unit 185 has been updated (in S11). Y) The morpheme analysis of all patent documents held in the patent database 200 is executed, and the morpheme information held in the morpheme information holding unit 181 is updated (S12). The dependency analysis unit 122 executes dependency analysis of all patent documents, and updates the dependency information held in the dependency information holding unit 182 (S13).

  In addition, when a patent document is added to the patent database 200 (Y in S14) or when an existing patent document is updated (Y in S15), the morpheme analysis unit 121 displays the added / updated patent document as a morpheme. The analysis is executed and the morpheme information is updated (S16). The dependency analysis unit 122 executes dependency analysis of the updated patent document and updates dependency information (S17).

  FIG. 6 is a flowchart showing a series of processing relating to extraction / aggregation processing in the document analysis support apparatus 100. The patent document search unit 130 searches for patent documents (S20). The analysis information extraction unit 150 extracts an analysis result (morpheme information and dependency information) about the patent document included in the search result (S21). The totaling unit 160 totals the analysis results under arbitrary conditions (S22). The display control unit 170 causes the user terminal 300 to display the counting results and the like (S23).

  According to the document analysis support apparatus 100 according to the present embodiment, the analysis results for all patent documents held in the patent database 200 are held. For this reason, it is not necessary to perform analysis at the time of analysis, and a desired analysis result may be extracted from the analysis results held in advance and aggregated. This makes it possible to analyze the patent document in a relatively short time.

(Second Embodiment)
The main difference between the document analysis support apparatus according to the first embodiment and the document analysis support apparatus according to the second embodiment is that each member when the patent document held in the patent database 200 is added / updated. Is the operation.
The document analysis support apparatus 100 according to the second embodiment has the same configuration as that in FIG. 2 except that it further includes a search condition holding unit. Hereinafter, the difference from the first embodiment will be mainly described.

The search condition holding unit holds the search conditions received by the search condition acquisition unit 131.
When the patent document held by the patent database 200 is updated, the patent document search unit 130 searches for the updated patent document based on the search condition held by the search condition holding unit. Specifically, first, the search condition acquisition unit 131 acquires a search condition from the search condition holding unit. The search formula generation unit 132 generates the search formula 21 based on this search condition. The search execution unit 133 searches the patent database 200 based on the search formula 21 generated by the search formula generation unit 132. At this time, the search execution unit 133 searches the patent documents newly added after the previous search and the patent documents updated after the previous search by referring to the update date and time of the patent documents. These processes may be scheduled to be executed after the patent database 200 is regularly updated.

  The patent document list acquisition unit 140 acquires a list of patent documents obtained as a result of the search by the search execution unit 133. This list is referred to as a patent document list 22 '. That is, the patent document list 22 'includes the document IDs of the added and updated patent documents. The morpheme information extraction unit 151 extracts the morpheme information of the patent document specified by the document ID 24 included in the patent document list 22 ′, and updates the extracted morpheme data holding unit 183. When the morpheme analysis of the updated patent document by the morpheme analysis unit 121 is not completed, the morpheme information extraction unit 151 waits for the completion before extracting the morpheme information.

  The dependency information extraction unit 152 extracts the dependency information of the patent document identified by the document ID 24 included in the patent document list 22 ′, and updates the extraction dependency data holding unit 184. If the dependency analysis of the updated patent document by the dependency analysis unit 122 is not completed, the dependency information extraction unit 152 waits for the completion and extracts the dependency information.

  The aggregation unit 160 performs the aggregation process again with the data held in the updated extracted morpheme data holding unit 183 and the extracted dependency data holding unit 184. When a patent document held in the patent database 200 is added / updated, a series of processing from the patent document search unit 130 to the analysis information extraction unit 150 is automatically performed.

  According to the document analysis support system 10 according to the present embodiment, the same operational effects as in the first embodiment can be achieved. In addition, according to the document analysis support system 10 according to the present embodiment, every time the patent database 200 is updated, a search is automatically performed based on a pre-registered search condition, and morphological analysis information to be analyzed and related The received information and the total result are updated. Thereby, the latest analysis result can always be obtained.

  The present invention has been described based on the embodiments. This embodiment is an exemplification, and it will be understood by those skilled in the art that various modifications can be made to combinations of the respective constituent elements and processing processes, and such modifications are also within the scope of the present invention. is there. A modification is shown below.

  A first modification will be described. In the first and second embodiments, the case where the document data is a patent document has been described, but the present invention is not limited to this. For example, the document data may be academic papers, newspapers, magazines, and other documents.

A second modification will be described. In the first and second embodiments, the case has been described in which the morpheme information held in the morpheme information holding unit 181 is configured to be one record for each morpheme, but is not limited thereto. For example, the morpheme information may be configured to be one record for each patent document. Specifically, for example, the data of each field of each record of FIG. It is also possible to configure such that one record is connected for each patent document. Thereby, since the number of records of morpheme information decreases, the search speed at the time of searching this improves. As a result, the speed when the morpheme information extraction unit 151 extracts the morpheme information of the patent document to be analyzed from the morpheme information holding unit 181 is improved.
The same applies to the dependency information held in the dependency information holding unit 182.

  A third modification will be described. Although not mentioned in the first and second embodiments, the document analysis support apparatus 100 includes a template that holds a template in which the aggregation unit 160 sets the items to be analyzed and the analysis results are aggregated. You may further provide a holding | maintenance part. As an example, the template includes: “The morpheme totaling unit 161 totals the appearance frequency of each morpheme in the morpheme information while grouping by age and applicant, and the dependency totaling unit 162 groups the morpheme by applicant. “Aggregating frequency of appearance of morphemes in dependency relationship with“ issue ”” is set. The template holding unit may hold a plurality of such templates, and the user only has to select a desired template. As a result, it is not necessary for the user to set which items are grouped and aggregated, and the burden on the user is reduced.

  A fourth modification will be described. In the first and second embodiments, the document analysis support apparatus 100 includes an analysis unit 120, a patent document search unit 130, a patent document list acquisition unit 140, an analysis information extraction unit 150, a totaling unit 160, and a display. Control unit 170, morpheme information holding unit 181, dependency information holding unit 182, extracted morpheme data holding unit 183, extraction dependency data holding unit 184, dictionary data holding unit 185, and morpheme total data holding unit 186 However, the present invention is not limited to this, and a part of the function of the document analysis support apparatus 100 may be transferred to another apparatus. For example, a patent document search device may be provided separately from the document analysis support device, and the function of the patent document search unit 130 may be provided thereto. Further, for example, an analysis device may be provided separately from the document analysis support device, and the function of the analysis unit 120 may be provided thereto.

  Any combination of the above-described embodiments and modifications is also useful as an embodiment of the present invention. The new embodiment generated by the combination has the effects of the combined embodiment and the modified examples. In addition, it should be understood by those skilled in the art that the functions to be fulfilled by the constituent elements described in the claims are realized by the individual constituent elements shown in the embodiments and the modified examples or by their linkage. .

  100 document analysis support device, 120 analysis unit, 121 morpheme analysis unit, 122 dependency analysis unit, 130 patent document search unit, 131 search condition acquisition unit, 132 search expression generation unit, 133 search execution unit, 140 patent document list acquisition unit 150 analysis information extraction unit, 151 morpheme information extraction unit, 152 dependency information extraction unit, 160 tabulation unit, 161 morpheme tabulation unit, 162 dependency tabulation unit, 170 display control unit, 181 morpheme information holding unit, 182 dependency information Holding unit, 183 extracted morpheme data holding unit, 184 extraction dependency data holding unit, 185 dictionary data holding unit, 200 patent database, 300 user terminal.

Claims (5)

An analysis unit that performs morphological analysis on a plurality of document data held in a database, and generates morpheme information obtained by dividing each of the plurality of document data into morphemes;
A morpheme information holding unit for holding morpheme information generated by the analysis unit ;
A search condition for searching for document data is input, and a search unit for searching for document data under the search condition;
A list acquisition unit for acquiring a list for acquiring a list of document data obtained as a result of the search by the search unit , among the document data held in the database;
An analysis information extraction unit that extracts morpheme information of document data included in the list from the morpheme information holding unit;
A totaling unit that totalizes the morpheme information extracted by the analysis information extraction unit ,
By the morphological analysis of the document data held in the database by the analysis unit, the morpheme information holding unit holds morpheme information of all document data targeted by the system,
When the document data held in the database is updated, the analysis unit performs a morphological analysis of the updated document data, updates the morpheme information,
The document collection support system , wherein the counting unit groups each morpheme included in the morpheme information by an item designated by the template, and totals the appearance frequency . The analysis unit further performs dependency analysis for generating dependency information between morphemes included in the morpheme information, and when the document data held in the database is updated, dependency analysis of the updated document data is performed. To update the dependency information,
A dependency information holding unit for holding dependency information generated by the analysis unit ;
The document analysis support system according to claim 1, wherein the analysis information extraction unit extracts dependency information of document data included in the list from the dependency information holding unit. Further comprising a dictionary data storage unit storing the dictionary data to be used for morphological analysis by pre-Symbol analyzer,
When the dictionary data is updated, the analysis unit performs morphological analysis and dependency analysis on a plurality of document data held in the database, and updates the morpheme information and the dependency information. The document analysis support system according to claim 2 . A search condition holding unit for holding a search condition for searching for document data;
When the document data held in the database is updated, the search unit searches the updated document data for document data that matches a search condition held in the search condition holding unit. The document analysis support system according to claim 3. An analysis function for performing morphological analysis on a plurality of document data stored in a database and generating morpheme information obtained by dividing each of the plurality of document data into morphemes;
A holding function for holding the generated morpheme information;
A search function for searching for document data is input, and a search function for searching for document data under the search condition;
A list acquisition function for acquiring a list of document data obtained as a result of search among the document data held in the database;
An analysis information extraction function for extracting morpheme information of document data included in the list;
The computer implements a tabulation function that tabulates the extracted morpheme information ,
By the morphological analysis of the document data held in the database by the analysis function, the holding function holds morpheme information of all document data targeted by the system,
When the document data stored in the database is updated, the analysis function performs morphological analysis of the updated document data, updates the morpheme information,
The said totaling function groups each morpheme contained in morpheme information by the item designated with the template, and totalizes the appearance frequency, The computer program characterized by the above-mentioned .

Images