JP2010218020A - Document management device, document management method, and program - Google Patents

Abstract

When registering a document in which attribute information characterizing the document is described in a database, the document and the attribute information are registered in association with each other without inputting the attribute information. Provided are a document management apparatus, a document management method, and a program that can be performed.
A document input unit of a document management apparatus inputs a patent publication file. The input document type determination unit 104 determines the file type of the patent publication file 100. The attribute extraction unit 106 extracts the publication number as the attribute information of the patent publication file 100 based on the attribute information specifying keyword indicated by the brackets in the patent publication file 100. The document information registration unit 108 registers the publication number and the patent publication file 100 in the database 112 in association with each other.
[Selection] Figure 1

Description

  The present invention relates to a document management apparatus, a document management method, and a program for managing a document and attribute information characterizing the document as mutually related data using an electronic computer, and in particular, features the document inside the document. The present invention relates to a document management apparatus, a document management method, and a program capable of registering an input document and the attribute information without inputting the attribute information when a document in which attribute information to be attached is input.

  In general, when data is managed using a database, each data is managed by a database schema including a data table in which a set of data item names representing the names of the data is defined. When registering data in the database, the input data is shaped from the input data into a form that matches the data table, a set of data corresponding to the data table is created, and registered in the database. Therefore, when inputting data, it is necessary to input data corresponding to the data item name together with the data item name.

FIG. 8 shows a conventional data management apparatus. The data management device includes data registration means, a database, and a database schema. Input data is input to the data registration means.
FIG. 9 shows an example of the input data shown in FIG. In the input data shown in FIG. 9, the data item name and the content of the data corresponding to the data item name are described for each line. For example, when registering patent gazette information, the application number, the publication number, A registered patent number and a gazette text position indicating a location on a storage area where a patent gazette text in which a full patent gazette is described are stored are described.

  FIG. 10 shows, as an example of the data schema shown in FIG. 8, a patent publication schema used when patent publication information is managed by a database. The patent publication schema is composed of two data tables. One is a number table in which application numbers, publication numbers, and registered patent numbers are defined. The other is a publication text table, in which a publication number and publication text storage position information are defined. A data definition language is used to define the data table.

  The data registration means forms the contents described in the input data into a form according to the data table and registers it in the database. The data management apparatus shown in FIG. 8 processes the input data line by line from the input data shown in FIG. 9, and extracts the data item name and the data information corresponding to the data item name. A set of data ((application number, publication number, registered patent number) and (publication number, gazette text position)) according to the table shown is created, and the set of data is registered in the database.

  In the data schema shown in FIG. 10, the public number is defined as a common item in two tables. As a result, the publication number is shared in both the number table and the gazette text table, and when the patent information is inquired to the database based on the publication number, the application number, the registered patent number, and the gazette text position are collectively displayed together with the publication number. It becomes possible to extract.

  However, in the case of such a data management apparatus, there has been a problem of erroneous registration of data due to an input error that occurs when data is input. In other words, if "publication number" is described in the place where "public number" should be described in the data item name, data registration to the database is impossible or erroneous registration (data item name mismatch problem) In addition, in the data contents, when the data of the publication number is erroneously described in the data portion corresponding to the public number, there is a problem that incorrect information is registered in the database (data content mismatch problem). In addition, when registering a plurality of data, it is necessary to create the data item name shown in FIG. 9 and data content information corresponding to the data item name every time, and there is a problem of complicated registration data creation. .

  Japanese Patent Application Laid-Open No. 2004-151867 discloses a method for collectively registering the same data to solve the problem of data registration complexity and erroneous data registration that occur when data of the same content is input. In Patent Document 1, when registering data having the same content, instead of sequentially inputting data, data according to a search condition is extracted from data registered in advance in a database, and the data according to the search condition is extracted. By performing the duplication processing, it is possible to avoid the trouble of manually inputting the same data and the problem of erroneous input.

JP 2007-86862 A

The problem of erroneous registration of input data and the problem of complexity of input work are important problems when registering data in the data registration apparatus. Has solved.
However, the problem of erroneous input of data input that occurs during data registration in a normal data registration apparatus has not yet been solved. Further, in Patent Document 1, it is necessary to input a data item name accompanying data input, and the problem of complicated input of the data item name has not been solved. Since the data item name is important information that links all the related information, if the data item name contains an error, the reliability of the data in the database is significantly impaired.

In addition, as shown in FIG. 9, when bibliographic information such as patent gazette text and application number, publication number, etc. is input, the bibliographic information is stored as data even though the bibliographic information is described in the patent gazette text. There was a problem of complexity because it was necessary to input to the registration means. Further, for example, when the public number is registered by mistake, there is a problem of erroneous registration in which the input public number is different from the public number described in the patent gazette text input together with the public number.
Therefore, the present invention has been made paying attention to the above-mentioned conventional unsolved problems, and when registering the document in which attribute information characterizing the document is described in the document, the attribute is registered. It is an object of the present invention to provide a document management apparatus, a document management method, and a program capable of associating and registering the document and the attribute information without inputting information.

  In order to solve the above problem, a document management apparatus according to claim 1 of the present invention includes a document input unit that inputs the document in which attribute information characterizing the document is described, and the document. In a document management apparatus comprising a database that manages the attribute information in association with each other, the attribute information extraction unit that extracts the attribute information from the document input by the document input unit, and the attribute information extraction unit A document information registration unit that associates the attribute information with the input document and registers it in the database, and the attribute information extraction unit is a keyword described in the input document, Based on an attribute information specifying keyword that specifies a location where attribute information is described, or an expression pattern described in the input document Based on the appearance frequency of the character string corresponding to the attribute information expression pattern representing said attribute information, and extracts the attribute information.

  According to the first aspect of the present invention, the document management device is based on an attribute information specifying keyword that specifies a location where the attribute information is described, which is described in the input document, or the The attribute information is extracted based on the appearance frequency of the character string corresponding to the attribute information expression pattern expressing the attribute information, and the extracted attribute information and the input document are associated and registered in the database. Therefore, the attribute information and the document can be associated and registered in the database simply by inputting the document to the document management apparatus. Therefore, the user does not need to specify or input the attribute information of the document, so that the registration becomes simple and the erroneous registration of the attribute information can be prevented.

  The document management apparatus according to claim 2 further includes an area estimation parameter management unit that manages area estimation parameters for estimating an appearance area in which the character string appears in the document. When the attribute information extraction unit extracts the attribute information based on the appearance frequency of the character string, the attribute information extraction unit estimates the appearance region using the region estimation parameter managed by the region estimation parameter management unit, and the estimation is performed. The appearance frequency of the character string is determined by searching the character string from the appearance region.

  According to the invention of claim 2, by estimating the appearance area using the area estimation parameter managed by the area estimation parameter management unit, and searching the character string from the estimated appearance area, In order to determine the appearance frequency of the character string, it is possible to increase the appearance frequency of the character string by performing a search limited to the appearance region where the character string representing the attribute information is likely to appear, It is possible to perform attribute extraction with high accuracy.

  The document management method according to claim 3, wherein the document management method is executed by a document management apparatus that manages a database by associating a document with attribute information characterizing the document and managing the document in which the attribute information is described. A document input step for inputting, and an attribute information specifying keyword for specifying a location in which the attribute information is described in the input document, the attribute based on the attribute information specifying keyword If the attribute information specifying keyword is not described in the input document and an attribute information expression pattern expressing the attribute information is described in the input document, the attribute information expression Extracted from the attribute information extraction step for extracting the attribute information based on the appearance frequency of the character string corresponding to the pattern, and the attribute information extraction step In association with a document whose serial was attribute information and said input, characterized in that it comprises a document information registration step of registering in the database.

  According to the third aspect of the present invention, when the attribute information specifying keyword for specifying the location where the attribute information is described is described in the input document, the document management apparatus The attribute information is extracted based on the attribute information specific keyword, and the attribute information specific keyword is not described in the input document, and an attribute information expression pattern expressing the attribute information is described. The attribute information is extracted based on the appearance frequency of the character string corresponding to the attribute information expression pattern, and the extracted attribute information and the input document are associated with each other and registered in the database. Therefore, without inputting the attribute information of the document, the attribute information and the document are associated with each other and registered and managed in the database only by inputting the document. It can become.

  The document management method according to claim 4, further comprising an area estimation parameter setting step for setting an area estimation parameter for estimating an appearance area in which the character string appears in the input document. In the attribute information extraction step, when the attribute information is extracted based on the appearance frequency of the character string, the appearance region is estimated using the region estimation parameter set in the region estimation parameter setting step, The appearance frequency of the character string is determined by searching the character string from the estimated appearance area.

  According to the fourth aspect of the present invention, in the attribute information extraction step, when the attribute information is extracted based on the appearance frequency of the character string, the region estimation parameter set in the region estimation parameter setting step is The appearance region is estimated using the character string, and the character string representing the attribute information is likely to appear in order to determine the appearance frequency of the character string by searching the character string from the estimated appearance region. By performing the search limited to the appearance area, the appearance frequency of the character string can be increased, and the attribute extraction with high accuracy can be performed.

  The program according to claim 5 can be stored in a computer and executed by being downloaded from a server or copied from a recording medium, whereby the method according to claim 3 or 4 can be realized by the computer. .

  According to the present invention, the document management apparatus is configured to use an attribute information specifying keyword that specifies a location in which the attribute information is described in the input document, or an attribute that represents the attribute information. In order to extract the attribute information based on the appearance frequency of the character string corresponding to the information expression pattern, and to associate the extracted attribute information with the input document and register them in the database. The attribute information and the document can be associated with each other and registered in the database simply by inputting to the document management apparatus. Therefore, the user does not need to specify or input the attribute information of the document, so that the registration becomes simple and the erroneous registration of the attribute information can be prevented.

It is a figure which shows the structure of the document management apparatus which concerns on Example 1 of this invention. It is a flowchart which shows the flow of the determination method which the input document classification determination part with which the document management apparatus concerning Example 1 of this invention is provided performs. It is a figure which shows an example of the database schema with which the document management apparatus which concerns on Example 1 of this invention is provided. It is a figure which shows the structure of the document management apparatus which concerns on Example 2 of this invention. It is a figure which shows an example of the publication number format pattern of the patent gazette file input into the document management apparatus which concerns on Example 2 of this invention. It is a figure which shows the structure of the document management apparatus which concerns on Example 3 of this invention. It is a flowchart which shows the flow of the determination method of the appearance frequency of the character string which the attribute extraction part which concerns on Example 3 of this invention performs. It is a figure for demonstrating the conventional general data registration method. It is a figure which shows an example of patent information input data. It is a figure which shows an example of the database schema in the case of managing patent information.

  Hereinafter, as an embodiment of the present invention, a case where a document is a patent publication file 100 will be described with reference to the drawings. In the drawings referred to in the following description, the same parts as those in the other drawings are denoted by the same reference numerals.

  FIG. 1 is a configuration diagram of a document management apparatus according to the first embodiment. As shown in the figure, the document management apparatus of the present invention includes a document input unit 102, an input document type determination unit 104, an attribute extraction unit 106, a document information registration unit 108, a database schema 110, and a database 112. These functions are realized by a CPU (Central Processing Unit) (not shown) included in the document management apparatus executing a program stored in a storage device such as a hard disk or a ROM (Read Only Memory).

The patent publication file 100 is input to the document input unit 102. Here, the patent gazette file 100 is an electronic version patent gazette text file or an electronic version patent gazette pdf file publicly disclosed by IPDL (Patent Electronic Library).
In this embodiment, a case where a patent publication file 100 described in a text format is input is shown.
The input document type determination unit 104 determines the file type of the input patent gazette file 100 and determines whether the patent gazette file 100 is a text file or a pdf file. Then, the determination result is set in a storage area for storing the input document type, which is provided in advance in the storage device.

  FIG. 2 is a flowchart showing the flow of the determination method performed by the input document type determination unit 104. First, the input document type determination unit 104 starts processing in S200, and proceeds to S210. In S210, it is determined whether or not the file type determination tool can be used. The file type determination tool is not prepared in advance in the document management apparatus, but refers to a tool prepared in addition to the document management apparatus used by the document management apparatus. A typical file type determination tool includes a file command attached to UNIX (registered trademark) and OS (operating system). If the file type determination tool can be used, the process proceeds to S230, and the file type determination tool is caused to read the patent publication file 100. The determination result of the file type determination tool is set to the input document type that is the final result of the input document type determination unit 104, and the process proceeds to S260 and the process is terminated.

  If it is determined in S210 that the file type determination tool cannot be used, the process proceeds to S220. In S220, it is determined whether or not the file name extension of the patent publication file 100 is “pdf”. If the extension of the file name is “pdf”, the process proceeds to S240. In S240, the input document type is set to “pdf”, the process proceeds to S260, and the process ends.

If the extension of the file name is not “pdf” in S220, the process proceeds to S250. In S250, it is determined that the patent publication file 100 is a text file, and the input document type is set to “txt”. Thereafter, the process proceeds to S260 and the process is terminated.
The attribute extraction unit 106 extracts attribute information described in the patent publication file 100 from the patent publication file 100 based on the input document type obtained by the input document type determination section 104.

In this embodiment, “application number”, “publication number”, and “invention name” are extracted as attribute information.
When the input document type is “txt”, the entire text including bibliographic information is written in a text format that can be processed mechanically, and “application number”, “publication number”, and “invention name” are respectively It is possible to extract the application number, the publication number, and the name of the invention with the brackets in the corners as the attribute information specifying keyword. For each of the attribute information, the attribute extraction unit 106 extracts, as the attribute information, a portion from the end of the attribute information specific keyword corresponding to the attribute information to the end of the sentence with a line feed code.

The document information registration unit 108 includes the database shown in FIG. 3 defined in advance in the database schema 110 from the attribute information of “application number”, “publication number”, and “invention name” extracted by the attribute extraction unit 106. Data information according to the table is created and registered in the database 112.
As shown in the present embodiment, when registering the patent gazette file 100 in the database 112, the user inputs “application number”, “publication number”, and “invention name” which are attribute information of the patent gazette file 100. Therefore, it is not necessary to designate and input separately from the patent publication file, and the present invention makes it possible to realize a document management apparatus that simplifies the registration process and does not cause the problem of erroneous registration of the attribute information.

FIG. 4 is a configuration diagram of the document management apparatus according to the second embodiment.
The document management apparatus according to the second embodiment is different from the document management apparatus according to the first embodiment in that the type of the patent publication file 100 is changed to a pdf file and the processing contents of the attribute extraction unit 206 are the same as those in the first embodiment. Is different. Other than that is the same as Example 1.
In the patent publication file 100 in the pdf format, bibliographic information is inserted as image information, and even when the pdf file is converted into text information that can be processed mechanically, the bibliographic information can be processed mechanically. It cannot be extracted as information.

  In the patent publication file 100 in the pdf format, a public number is described in the header portion of each page. In the document management apparatus according to the present embodiment, when the patent publication file 100 in the pdf format is input, the public number described in the header part is extracted as attribute information and corresponds to the patent publication file table of the database schema 110. Data is registered in the database 112.

  Specifically, the attribute extraction unit 206 reads text information obtained by converting the patent publication file 100 in the pdf format from the beginning to the end, and obtains a character string according to the public number format pattern according to FIG. FIG. 5 shows a public number format pattern using regular expressions. FIG. 5A shows a pattern that begins with “JP,” followed by either “Hira” or “Akira,” followed by a numeric string, “−”, and a numeric string. Represents. FIG. 5B shows a pattern starting with “JP,” followed by a number string composed of four numbers, “−”, and a number string. The attribute extraction unit 206 determines the character string having the highest appearance frequency as the publication number of the patent publication file 100 from the obtained character string and the appearance frequency of the character string.

  The patent publication file 100 includes the publication number of the prior art document in addition to the publication number of the patent publication file 100, and the attribute extraction unit 206 also extracts the publication number of the prior art document. However, the publication number of the patent publication file 100 is described in the header part of each page. Generally, the publication number of the patent publication file 100 is higher in appearance frequency than the publication number of the prior art document, and the appearance frequency is higher. By selecting a column, the attribute extraction unit 206 can correctly extract the publication number of the patent publication file 100.

FIG. 6 is a configuration diagram of the document management apparatus according to the third embodiment.
The third embodiment is different from the second embodiment in that an operation content of the attribute extraction unit 306 and a region estimation parameter management unit 314 are added. Other than that, it is equivalent to the second embodiment.
The area estimation parameter management unit 314 manages a value corresponding to the number of lines described on one page of the patent publication file 100. In this embodiment, the initial value of the value corresponding to the number of rows is set to “58”.
The attribute extraction unit 306 does not perform a search process on the entire text of the patent publication file 100 as in the attribute extraction unit 206 according to the second embodiment, and only for a specific appearance region that is estimated that a public number appears. The difference is that a character string corresponding to the public number format pattern shown in FIG. 5 is searched.

  This is because, in the attribute extraction unit 206 according to the second embodiment, the entire text of the input patent gazette file 100 is used as a character string search area. Therefore, if the patent gazette file 100 is short, the publication number of the prior art document is set. This is because the appearance frequency and the appearance frequency of the public number may be equal, and in this case, there are a plurality of character strings having a high appearance frequency, and there is a problem that the public number cannot be extracted. On the other hand, in the attribute extraction unit 306 according to the present embodiment, the character string having a high appearance frequency can be obtained by limiting the search area for the character string to an appearance area where the public number is likely to appear. Solves multiple existing problems.

  When the patent publication file 100 is a pdf file, a public number is described in the header portion of each page, and this public number is extracted as attribute information characterizing the input file. The public number described in the header part appears almost regularly with the number of lines in one page separated. In the present embodiment, the problem of the presence of a plurality of character strings having a high appearance frequency is solved by specifying and searching the area of the public number that appears regularly as an appearance area.

FIG. 7 is a flowchart showing the flow of the method for determining the appearance frequency of the character string executed by the attribute extraction unit 306. The attribute extraction unit 306 first starts processing in S1000, and proceeds to S1010.
In S1010, a search area for the search to be performed first is set, and parameters used in the entire process are initialized. Here, start_position and end_position indicate a search start position and a search end position that designate the search area. “end_position = −1” indicates that the search area extends to the end of the input patent gazette file 100. previous_found_position represents the position of the character string found in the previous search, and areasize represents half the size of the search area. Here, the position of the character string represents the number of lines from the beginning of the patent publication file 100 in which the character string exists.

In S1020, a character string corresponding to the public number format pattern shown in FIG. 5 is searched from the determined search area. The search starts from the top of the search area. As soon as the character string is found, the process in S1020 is terminated, and the process proceeds to S1030.
In S1030, it is determined whether the character string is found. If not found, the process moves to S1040. If found, the process moves to S1070.

In S1070, the position of the found character string is set in found_position, 1 is added to the appearance frequency of the found character string, and the process proceeds to S1080.
In S1080, it is determined whether or not start_position is 0, and it is determined whether or not the found character string is the first found character string. When Start_position is 0, it is determined that the character string is found first, and the process proceeds to S1100. If Start_position is not 0, the process moves to S1090.

  In S1090, the distance from the location where the character string corresponding to the public number format pattern described in FIG. 7 was found last time to the location found this time is found from found_position-previous_found_position, and the page row number data indicating the distance is stored in the area. The estimated parameter management unit 314 is notified, and the process proceeds to S1100. The area estimation parameter management unit 314 manages page line number data indicating the distance between character strings corresponding to the public number format pattern shown in FIG.

In S1100, the value of found_position is set in previous_found_position for the next search, and the process proceeds to S1110.
In S1110, the page line number data is acquired from the area estimation parameter management unit 314, set to numlines, and the process proceeds to S1120.
In S1120, a search area for the next search is designated, and the process proceeds to S1020. The search area is determined by start_position = found_position + numlines−areasize and end_position = found_position + numlines + areasize. This is because an area of the front and back areasize centering on a position away from numlines from the place where the character string corresponding to the public number format pattern described in FIG. 5 was found last time is an appearance area where the character string is likely to appear. This indicates that the appearance area is set as the search area.

  On the other hand, if no character string is found in the search area in S1030, it is checked in S1040 whether end_position is −1, and it is determined whether the entire patent publication file 100 has been read. If end_position is −1, it is determined that the entire patent publication file 100 has been read, the process proceeds to S1050, and the process of the attribute extraction unit 306 is terminated. If end_position is not -1, the end of the search area is set to the end of the patent publication file 100, and the process proceeds to S1020 to repeat the process.

  As described above, since the document management apparatus can automatically extract attribute information such as a publication number described in the inputted patent gazette file 100, the patent gazette file 100 is stored in the document management apparatus. It is possible to register and manage the patent gazette file 100 and the attribute information in the database 112 simply by inputting them. Therefore, the user does not need to specify or input attribute information, and registration is simplified and erroneous registration of attribute information can be prevented.

In addition, regarding the method for extracting attribute information, the methods described in the first and second embodiments or the first and third embodiments can be combined. For example, when an attribute information specifying keyword for specifying a location where attribute information is described is described in the patent publication file 100, a publication number is extracted based on the attribute information specifying keyword. When the attribute information specific keyword is not described in the gazette file 100 and an attribute information expression pattern such as a public number format pattern is described, it is based on the appearance frequency of the character string corresponding to the attribute information expression pattern. Attribute information may be extracted.
In the above-described embodiment, an example in which the document is the patent publication file 100 has been described. However, the document is not limited to the patent publication file 100, and examples thereof include contracts, application forms, medical records, and collections of papers. Any document that has predetermined attribute information described in the document and should be managed in association with the predetermined attribute information is conceivable.

100 Patent Gazette File 102 Document Input Unit 104 Input Document Type Determination Unit 106 Attribute Extraction Unit 108 Document Information Registration Unit 110 Database Schema 112 Database 206 Attribute Extraction Unit 306 Attribute Extraction Unit 314 Area Estimation Parameter Management Unit

Claims (5)

In a document management apparatus comprising: a document input unit that inputs the document in which attribute information characterizing the document is described inside the document; and a database that manages the document in association with the attribute information.
An attribute information extraction unit that extracts the attribute information from the document input by the document input unit;
A document information registration unit that associates the attribute information extracted by the attribute information extraction unit with the input document and registers it in the database;
The attribute information extraction unit
Based on an attribute information specifying keyword that specifies a location in which the attribute information is described, which is a keyword described in the input document, or
Based on the appearance frequency of the character string corresponding to the attribute information expression pattern that is the expression pattern described in the input document and expresses the attribute information,
A document management apparatus that extracts the attribute information. An area estimation parameter management unit for managing area estimation parameters for estimating an appearance area in which the character string appears in the document;
The attribute information extraction unit
When extracting the attribute information based on the appearance frequency of the character string,
Determining the appearance frequency of the character string by estimating the appearance region using the region estimation parameter managed by the region estimation parameter management unit and searching the character string from the estimated appearance region The document management apparatus according to claim 1. In a document management method executed by a document management apparatus that associates a document with attribute information characterizing the document and manages it in a database,
A document input step for inputting the document in which the attribute information is described;
When an attribute information specifying keyword that specifies a location where the attribute information is described is described in the input document, the attribute information is extracted based on the attribute information specifying keyword,
If the attribute information specific keyword is not described in the input document and an attribute information expression pattern expressing the attribute information is described, characters corresponding to the attribute information expression pattern An attribute information extraction step of extracting the attribute information based on the appearance frequency of the column;
A document management method comprising: a document information registration step of registering the attribute information extracted from the attribute information extraction step and the input document in association with each other in the database. An area estimation parameter setting step for setting an area estimation parameter for estimating an appearance area in which the character string appears in the input document;
In the attribute information extraction step, when the attribute information is extracted based on the appearance frequency of the character string, the appearance region is estimated using the region estimation parameter set in the region estimation parameter setting step, The document management method according to claim 3, wherein the appearance frequency of the character string is determined by searching the character string from the estimated appearance area.   The program for making a computer perform the method described in Claim 3 or 4.

Images