JP2008158918A - Document management keyword extraction system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a document management keyword extraction system for efficiently and accurately setting, in management of a document by connecting a management purpose thereof with a keyword contained in the document, a keyword matched to the management purpose. <P>SOLUTION: The system comprises a client computer 2, a server 1, and a database 3, in which a plurality of files to be managed with the same management purpose is stored in a certain holder (e.g., A), and a character string composed of continuous n pieces of characters is extracted for each appearance place in the files, wherein as n, a natural number can be set as n. Such extraction is performed to all files (b, c, and d) stored in the holder A, and if the character string and the appearance place thereof are common to all the files, this character string is extracted and recorded as a keyword matched to the management purpose of the holder A. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a document management keyword extraction system, and more particularly, a document management keyword extraction system capable of more specifically extracting keywords suitable for a management purpose from sentences to be managed according to the management purpose and efficiently managing a document using the keyword. About.

There is a case where a document to be managed is automatically selected according to a certain management purpose from a huge amount of electronic documents existing in the information system (see, for example, Patent Document 1).
JP-A-6-187373

  In this way, in the process of automatically selecting documents to be managed in accordance with a certain management purpose, when using a mechanism for setting a keyword and selecting a document that includes the keyword, this keyword is appropriate depending on the subjectivity of the individual. In some cases, the necessary documents cannot be reliably selected because the administrator does not know what keywords should be set.

  The present invention relates to a document for efficiently and correctly setting a keyword suitable for the management purpose when managing the document by linking the management purpose and a keyword included in the document when managing the document. The purpose is to provide a management keyword extraction system.

  The present invention includes a client computer, a server, and a database. A plurality of files managed for the same management purpose are stored in a folder, and n characters (for each appearance location in the plurality of files). n is a natural number), and for all the files stored in the folder, the continuous character string of the n characters is extracted for each appearance location in the plurality of files. And detecting the character string and its appearance location common to all the plurality of files, and extracting and recording the detected character string and its appearance location as keywords suitable for the management purpose of the folder It is a management keyword extraction system.

  Further, the present invention performs a character string search for the appearance location of the keyword for files other than the plurality of files stored in the folder, and detects a file in which the extracted keyword matches the appearance location. The document management keyword extraction system warns a matched file as a file to be managed for the purpose of managing the folder, and moves or copies the file to be managed to the folder.

  Then, the present invention provides a document management keyword extraction system for writing a warning when an extracted keyword is not present in the corresponding appearance location in the file newly added to the folder and warning the extracted keyword It is.

  Furthermore, the present invention is a document management keyword extraction system that issues a warning when a newly added file cannot be written as a read-only file.

  According to the document management keyword extraction system of the present invention, keywords have been set in many cases depending on personal experience so far. However, keywords are assigned to a plurality of files under the same management purpose. By combining the possible character strings with the appearance locations as keywords, more detailed keywords can be extracted, and keywords suitable for management purposes can be set efficiently and correctly.

The best mode for carrying out the present invention will be described.
An embodiment of the document management keyword extraction system of the present invention will be described with reference to the drawings. FIG. 1 is a system configuration diagram of a document management keyword extraction system according to an embodiment.

  As shown in FIG. 1, the document management keyword extraction system of this embodiment includes a plurality of client computers 2 and a server 1 on the manager or user 4 side. Each client computer 2 can store a file in the server 1. The server 1 can extract keywords common to all the files in the folder, and record the extracted keywords in the database 3. The server 1 issues a warning to the administrator or the user 4 when there is another file to be managed in the system or when there is no appropriate keyword in the file in the folder.

  In the document management keyword extraction system of the present embodiment, by executing the document management keyword extraction method, the keyword setting so far has often been set depending on personal experience etc., but for the same management purpose A character string that can be used as a keyword can be combined with the appearance location as a keyword, and more detailed keyword extraction can be performed.

  An example of a processing procedure in the document management keyword extraction system of this embodiment will be described. First, (1) a plurality of files managed for the same management purpose are stored in the same folder. FIG. 2A shows an example in which a plurality of files b, c, and d are stored in the folder A. Files b, c, and d are managed files that are managed for the same management purpose. Next, (2) a character string having n consecutive characters is extracted for each appearance location of each file. FIG. 2B shows an example of keywords extracted from the file b and their appearance locations.

  Next, (3) the procedure of (2) is performed on all files stored in the folder A. FIG. 2C shows an example of keywords and appearance locations in the file c. FIG. 2D shows an example of keywords and appearance locations in the file d.

  Next, (4) a character string common to all files in the folder A and its appearance location are extracted. FIG. 2E shows an example of character strings (keywords) common to all files b, c, and d in the folder A and their appearance locations.

  In this embodiment, when a document is managed, in a mechanism for managing the management purpose by linking the keyword included in the document file, it is possible to efficiently and more accurately and correctly set the keyword suitable for the purpose. it can. Specifically, n (n: natural number) document files suitable for a certain management purpose are put in one folder, a combination of a keyword and an appearance location common to all the document files is found, and Are registered as keywords suitable for management purposes. As shown in FIGS. 2 (A) to 2 (E), the appearance location is a header, metadata, footer, template, body (number of lines and number of characters in the lines). Point to.

  Here, an example of a procedure for extracting a character string common to all the files b, c, and d in the folder A and its appearance location in the embodiment will be described. FIG. 3 is an explanatory diagram of an extraction procedure in the document management keyword extraction system of the embodiment. The extraction procedure includes steps S1 to S4. In step S1, as illustrated in the process (1) of FIG. 2A, a management purpose is set, and document files b, c, and d that meet the management purpose are stored in one folder A. . In step S2, as exemplified in the process (2) in FIG. 2B, a character string (n characters continuous) is added to the file b in the folder A in the process (1) in FIG. Keyword) 50 and its appearance location 51 are extracted.

  In step S3, as shown in FIGS. 2C and 2D, each of the files c and d in the folder A in the process (1) of FIG. 2A is processed. That is, as shown in FIG. 2C, a character string (keyword) 52 of n characters continuous and its appearance location 53 are extracted, and a character string (keyword) of n characters continuous is extracted as shown in FIG. 54 and its appearance location 55 are extracted. In step S4, the extraction process is completed by extracting a common character string (keyword) 60 and its appearance location 61 for all the files b, c, and d in the folder A.

  Next, returning to FIG. 2F, (5) when the character string extracted in the process (4) exists in the file e in addition to the folder A, the administrator is notified.

  Next, (6) when a new file is generated / moved in the folder A, it is confirmed whether or not there is a keyword in the corresponding appearance location, and if there is no keyword in the corresponding appearance location, it is written. FIG. 2 (G) shows that when a new file is created / moved in folder A, it is confirmed whether or not there is a keyword at the corresponding appearance location. If there is no keyword at the corresponding appearance location, writing is performed. An example of reporting to the administrator when writing is not possible is shown.

  The procedure for confirming the existence of another file e in this embodiment will be described. FIG. 4 is an explanatory diagram of another file e existence confirmation processing procedure in the document management keyword extraction system of the embodiment. The existence confirmation procedure for the file e includes steps S10 to S16. In step S10, the character string (keyword) extracted by the process (4) in FIG. 2 (E) for the file e other than the folder A in the process (5) in FIG. Check if it exists.

  In step S11, when the character string (keyword) does not exist at the corresponding appearance location, the confirmation procedure is terminated. However, when the character string (keyword) exists at the corresponding appearance location, in step S12, In step S13, it is determined whether or not the file e is set to be moved or copied to the folder A. If the file e is not moved or copied, the confirmation procedure is terminated. When moving or copying, the file e is moved or copied in step S14. In step S15, if the file e can be moved or copied, the confirmation procedure ends. If the file e cannot be moved or copied, in step S16, the administrator is notified and the confirmation procedure is completed. To do.

  Next, an example of a procedure for writing a new file f in the embodiment will be described. FIG. 5 is an explanatory diagram of a procedure for writing to another file in the document management keyword extraction system of the embodiment. The procedure for writing to a new file in FIG. 5 includes steps S20 to S24. In step S20, the character string (keyword) in the process (4) of FIG. 2 (E) is added to the new file e generated or moved in the folder A in the process (6) of FIG. 2 (G). Check if it exists at the appropriate location.

  In step S21, when the character string (keyword) has a corresponding appearance location, the writing procedure is terminated. In step S22, if the character string (keyword) does not have the corresponding appearance location, the file f is a folder. When it is determined whether or not data can be written to A, and writing cannot be performed with the read-only file, the administrator is notified in step S23 and the writing procedure is terminated. When the file f can be written in the folder A, in step S24, a character string (keyword) is written at the corresponding appearance location, and the writing procedure is terminated.

Incidentally, in the embodiment, it is preferable to perform document management keyword extraction as follows.
[1] As a premise, a character string that does not become a keyword is created as a database. The character string that does not become a keyword is a character string that does not have a meaning composed of an instruction word described in a dictionary or the like, or hiragana or katakana. These character strings can be appropriately removed from the list of character strings that do not become keywords according to user settings. Also, a character string that does not become a keyword can be added according to user settings.

  [2] When the process (2) of FIG. 2B is performed, if a technique for extracting a natural sentence from a character string is used in combination, candidate keywords can be narrowed down, so that more reliable keyword extraction can be performed. .

  [3] The processes (1) to (6) in FIG. 2 are performed. When the process (2) in FIG. 2B and the process (3) in FIG. Check against the database of character strings that cannot be used as keyword candidates. By performing this operation, it is possible to narrow down the character string as a keyword to some extent.

  [4] When the process (2) in FIG. 2B is performed, it is possible to narrow down the keywords by decreasing the value of n in order from the largest number. The initial value of n is set in advance when the present invention is executed, but in practice, it is desirable to start with a value of about 8 to 10.

  [5] When the process (3) in FIG. 2C is performed, if there are too many files in the folder A, it is difficult to find a keyword common to all of them. On the other hand, if the number of files is too small, there may be too many candidate keywords. Therefore, in practice, preferably three to four files that are strictly suitable for management purposes are selected as illustrated in FIG. 2, and keywords appear only in limited places such as metadata, headers, and footers. , Gradually increase the number of files to expand the keyword candidate location.

  [6] The processes (1) to (4) in FIG. 2 are performed whenever the files in the folder A increase or decrease, and maintenance is performed so that an appropriate combination of keyword candidates and their locations can be maintained. desirable.

  [7] When the process (5) of FIG. 2 (F) is performed, it takes time to search if there are many files in the system. Therefore, the search is performed for all files at first, but from the next time, the search time is shortened by searching only for files that have been changed or added since the first execution. Can do.

  In the process (6) of FIG. 2G, when creating a new document file e, if the new document file e is created by copying a previously created document file, the metadata created before that is created. May be taken over, and in some cases, a customer may be given a name with the name of another customer. This is a phenomenon that occurs because the metadata is usually not conscious, but by using the present invention, when storing a new document file e in a folder, the specified metadata is forcibly written. Can avoid these problems.

  In the document management keyword extraction system of the embodiment, a plurality of files managed for the same management purpose are stored in a certain folder (for example, A), and a character string of n characters continuous for each appearance location in the file. To extract. n is a natural number that can be set for all files (b, c, and d in the example shown) stored in the folder A, and the character string and its appearance location are set in all the above files. If there is something in common, it is extracted and recorded as a keyword suitable for the management purpose of folder A. In the past, keyword settings were often set depending on personal experience, etc., but for multiple files under the same management purpose, character strings that can be used as keywords are combined with their appearance locations. Therefore, more detailed keywords can be extracted, and keywords suitable for management purposes can be set efficiently and correctly.

  Further, in the document management keyword extraction system of the embodiment, a character string search is performed for a keyword appearance location for a file other than the file stored in the folder A (referred to as file e in the illustrated example), If there is a file whose appearance matches the extracted keyword, the file e is also warned as a file to be managed for the management purpose of the folder A, and if possible, the file is moved or copied to the folder A. Is based on preset contents. As a result, even if the file is other than the file stored in the folder, the keyword suitable for the management purpose can be set efficiently and correctly.

  In the document management keyword extraction system of the embodiment, if the extracted keyword does not exist in the corresponding appearance location in the file newly added to the folder A (referred to as f in the illustrated example), the keyword is determined to be applicable. A warning is given when the file f is a read-only file and cannot be written. As a result, even if the file is newly added to the folder, keywords suitable for the management purpose can be set efficiently and correctly.

  Although the embodiments have been described above, the present invention is not limited to the above-described embodiments, and various modifications can be adopted without departing from the scope of the claims.

The system block diagram of the document management keyword extraction system of an Example. Explanatory drawing of the document management keyword extraction system of an Example. Explanatory drawing of the extraction procedure in the document management keyword extraction system of an Example. Explanatory drawing of the presence confirmation of the other file in the document management keyword extraction system of an Example. Explanatory drawing of the write-in procedure to the new file in the document management keyword extraction system of an Example.

Explanation of symbols

1 Administrator server 2 Client computer 3 Database 4 Administrator or user A Folders b, c, d Folder A files e Files other than folder A f New files

Claims (4)

  A client computer, a server, and a database are provided, a plurality of files managed for the same management purpose are stored in a folder, and n characters (n is a natural number) for each appearance location in the plurality of files For all the files stored in the folder, for each occurrence location in the plurality of files, the continuous character string of n characters is extracted, and the characters are extracted. Detecting a character string and its appearance location that are common to all of the plurality of files, and extracting and recording the detected character string and its appearance location as a keyword suitable for the management purpose of the folder Document management keyword extraction system.   For a file other than the plurality of files stored in the folder, perform a character string search for the appearance location of the keyword, detect a file that matches the extracted keyword and the appearance location, The document management keyword extraction system according to claim 1, wherein a warning is given as a file to be managed for the purpose of managing the folder, and the file to be managed is moved or copied to the folder.   3. The document management keyword extraction system according to claim 2, wherein when the extracted keyword is not present in the corresponding appearance location in the file newly added to the folder, the extracted keyword is described in the corresponding appearance location and a warning is given. .   4. The document management keyword extraction system according to claim 3, wherein a warning is issued when the newly added file is a read-only file and cannot be written.

Images