JP2005215717A - Document processor with security function - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To further effectively utilize document data without impairing information property of a document with enhanced security by masking a word included in the document data, the disclosure of which is to be restricted. <P>SOLUTION: The access level of a user is acquired by a user authentication means, and the disclosure-restricted word included in the document to be disclosed according to the access level is specified by collation of words included in the document with disclosure-restricted word data. A masking character string composed of a character string showing the kind of the disclosure-restricted word and an additive character string for discriminating a different word is generated, and the document is disclosed with the disclosure-restricted word being substituted with the masking character string. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a method and an apparatus for masking a word (replace a designated word with another character string) for restricting disclosure contained in a document to be disclosed and word information (such as word information extracted from the document).

  Companies are accumulating a lot of document data such as inquiry histories from customers and in-house documents, and there is an increasing need to make effective use of these document data. Document processing apparatuses such as a document search system and a text mining system are examples of apparatuses for effectively utilizing these document data. However, document data and document data, such as personal names and company names that are listed as proper nouns, are included in the document data because some users want to restrict their disclosure (hereinafter referred to as “public restriction words”). Users who can access the word information extracted from are limited to some users such as administrators.

  However, it is very important to make document data available to the extent that there is no problem, and it is very important to effectively use the document data, and a mechanism that can securely browse the document data has been desired. Therefore, as a document publishing method, a technique for publishing by replacing the restricted word with another character string has appeared. For example, Japanese Patent Application Laid-Open No. 11-143871 (Patent Document 1) uses a method of specifying a part whose publication is restricted according to a user profile and painting the part black. Japanese Laid-Open Patent Publication No. 2002-312362 (Patent Document 2) uses a method in which proper nouns are replaced with initial character strings and published.

Japanese Patent Laid-Open No. 11-143871

JP 2002-312362 A

  The above conventional technique has a problem that when the restricted word is masked with black (the designated portion is blacked out), the type of the word in the masking portion is not known and the content is difficult to grasp. For example, if you mask “Mr. Saito” in the sentence “I want to ask Mr. Saito because my computer has broken down,” “I want to ask you ■ because my computer has failed.” The part corresponding to “■” Names, personal names, product names, or other types of words cannot be distinguished, making it difficult to grasp the contents. In addition, in masking using a character string or initial character string whose type is known, it is easy to grasp the content because the type of the part indicated by the masking character string can be confirmed, but on the other hand, the masking character string in some documents If the indicated word is specified, there is a problem because the word indicated by the same masking character string included in another document is also specified. For example, if there are sentences such as “Mr. Sato was admitted” or “Mr. Sato was unkind”, masking that “[Employee S] was admitted” or “[Employee S] was unkind” A person who knows that he is hospitalized sees “Employee S” as “Mr. Sato” when viewing the document, and “Employee S” of “[Employee S] is unkind” is “Mr. Sato” "" Is specified. It may be possible to simply mask “[employee]”, but if there are multiple masking targets of the same type in the same document, whether each masking word indicates the same person or a different person There is a problem that it is difficult to grasp the contents without understanding.

An object of the present invention is to provide a masking character string whose type is known, and even if a word indicated by the masking character string is specified in some documents, the word indicated by the masking character string of another document is not specified and is the same Different words contained in a document are masked so that they can be distinguished.
Another object of the present invention is to securely disclose document information extracted from a document, and to provide a mechanism for retrieving a document from document information including a disclosure restriction word.

In order to achieve the above object, an outline of the invention disclosed in the present application is as follows.
The document processing apparatus with a security function according to the present invention specifies a public restriction word included in a document and word information extracted from the document (for example, a word pair having a dependency relationship), and identifies the word as a type and an additional character string. Replace with a masking string consisting of and publish. Here, the additional character string is generated randomly or based on a predetermined rule. In order to maintain the information property of documents, there are provided a feature that masking character strings for different words in the same document are different from each other and a masking character string for the same words in the same document are the same. For information security, a masking character string of the same word must be common to all documents. In addition, in a document search with word information specification, if the word information that is a search condition includes an open restriction word, the word information constituting the word information other than the open restriction word is searched or expanded to the same type of word. A screen that displays whether or not to search, and provides a mechanism that allows you to specify whether or not to perform the search.

Since the document processing apparatus of the present invention masks a different character string if it is a different document even if it is the same open restriction word, even if the word indicated by the masking character string is specified in some documents, There is an effect that the word indicated by the masking character string is not specified. Further, since the masking character strings of different words included in the same document are different, there is an effect that it is possible to prevent the contents from being misunderstood without impairing the information property of the document.
As another effect, even if the restricted word contained in the word information is masked, a document search with word information designation can be realized securely and easily.

  The purpose of minimizing information loss and securely publishing documents containing restricted words is realized by improving the masking string generation method.

A first embodiment of the present invention will be described below with reference to the drawings.
FIG. 1 is a configuration diagram of a document disclosure apparatus according to the present embodiment. This apparatus includes a central processing unit CPU 101, a main memory 102, a display device 103, an input device 104, and a storage device 110. The storage device 110 includes an OS (operating system) 111, document data 112, public restriction word data 113, user information data 114, user authentication program 115, document search program 116, and public document generation program 117. A document display program 118 is stored.

  In the public restriction word data 113, a word for performing public restriction is registered. In addition to the restricted word, the access level of the user who restricts the disclosure, the type of word, and the like are also registered corresponding to each word. In the user information data 114, authentication information and access level for each user are registered. The user authentication program 115 receives the user ID and password input from the input device 104, specifies the user by collating the input information with the user information data 114, and acquires the access level of the user. Further, the document search program 116 receives an input of a word from the input device 104 and extracts a document including the word from the document data 112. Further, the public document generation program 117 specifies a public restriction word included in a document to be published (hereinafter referred to as a public document) using the access level of the user and the public restriction word data 113, and masks the public restriction word. A public document is generated by replacing the public restriction word with the generated character string. In addition, the document display program 118 causes the display device 103 to display a public document. The above program is read into the main memory 102 and executed under the control of the CPU 101.

  Next, the processing flow of the present embodiment will be described with reference to the flowchart of FIG. First, the user authentication program 115 uses a user authentication means such as a user ID and a password, an IC card, and biometric authentication to identify a user who wants to view a document (S201), and obtains the access level of the user. For example, if user information is managed using the table (user information data 114) shown in FIG. 3, when the user ID “1002” and the password “gfddf” are input, the user is the user ID “1002”. The access level 2 is acquired. Other methods may be used as long as the user access level can be acquired.

  Next, the document search program 116 specifies a search condition such as a word included in the document, extracts a document that matches the search condition, and specifies a public document (S202). For example, if a document is managed by the table shown in FIG. 4 (public restriction word data 113) and “hard disk” is used as a search condition, the document with the document ID “10002” including “hard disk” becomes a public document. The public document can be identified by any method as long as the public document can be identified, such as designation of a document ID, designation of registration date and category attached to the document data, and designation of the full text. There may be a plurality of public documents. Next, the public document is read from a storage device such as a memory or a hard disk (S203), the words included in the public document are compared with the public restriction word data, and the public restriction word contained in the public document is specified (S205). . Here, when the user's access level value is smaller than the access level value set in the public restriction word, it is not set as the public restriction word. For example, if the public restriction word is managed in the table shown in FIG. 5 and the public document is “Mr. Yamada was unfriendly”, if the user is an access level 2, the public restriction word data includes “ “Yamada-san” is identified as a restricted word. When the same public document is displayed for the access level 1 user, “Mr. Yamada” included in the public restriction word data is not specified as a public restriction word and is displayed as it is without masking. Note that there may be a plurality of public restriction words in the same sentence. If there are multiple release restriction words in the same sentence, release restriction is performed for each.

Next, the public document creation program 117 acquires the type of the restricted word contained in the public document from the table shown in FIG. 5, and generates a masking character string composed of the type and the additional character string (S206). ). The additional character strings are randomly generated, and are generated so as to be different additional character strings when the documents are different even if they are the same open restriction word. Therefore, even if the same open restriction word is used, different additional character strings are generated if the documents are different. For example, a character string generated using a random number R (0 ≦ R <1) having a decimal value of “Employee” as the type of the restricted word “Mr. Yamada” is
int (R × 26) + 1 = 2 B (1 is A, 2 is B, 3 is C, and 26 is Z)
In this case, “Employee B”, which is a combination of the type “Employee” and the additional character string “B”, is the masking character string to be generated. However, if the generated masking character string is the same as the masking character string of another public restriction word already included in the same document, a masking character string is generated again, and different public restriction words in the same document are differently masked. Generate a string. Therefore, even if the masking character string is replaced, the masking character string of different open restriction words in the same document can be distinguished. In addition, if a masking character string is created separately for each document, it is possible to prevent the word before masking from being identified if the same masked part is compared in multiple documents. Information security can be improved.

  As the additional character string, any character string may be used as long as the character string can be distinguished, such as “001”, “AB”, and “◯”. In addition, color information is used instead of character strings, such as “Yamada-san” is displayed as “employee” with red text color, “Sato-san” is displayed as “employee” with blue text color. It doesn't matter. In addition, the additional character string is not generated randomly, but the alphabet from A to Z is used in order, for example, by using a method of extracting and using character strings in a predetermined order from a predetermined character string set. It doesn't matter. In addition, the character string indicating the type and the additional character string may be used as a masking character string only when the same document has a plurality of open restriction words of the same type, and only the type may be used as the masking character string in other cases.

  Next, the document display program 118 generates a public document by replacing the public restriction word included in the public document with the generated masking character string (S207), and displays the public document on a display device such as a personal computer display. (S208). When displaying, highlighting such as reversing the display color of the masking character string or enclosing it with a square, etc., and making it possible to visually recognize that these parts have been masked is convenient. improves. A user who has the authority to view all words can publish a document by skipping the steps S205 to S207 (S204). If a request to publish another document is received after the designated document is published, steps S202 to S208 are executed (S209).

  According to the above embodiment, different masking character strings are used for different documents even if they are the same open restriction word. This has the effect of ensuring document security without affecting the viewing of other documents. In addition, since the different restriction character strings included in the same document use different masking character strings so that each restriction word can be distinguished, there is an effect of preventing misunderstanding of the contents of the document.

  The second embodiment of the present invention provides a document processing apparatus capable of safely publishing document information such as word information extracted from a plurality of documents in addition to the document publishing apparatus described in the first embodiment. Here, the word information is a set of words composed of at least one word such as a word pair having a dependency relationship, but any method may be used as long as the word set can be extracted from the document data. Note that the word information is used for grasping the tendency of the contents of the document data. For example, when the number of word pairs “PC-failure” is large, the document data indicates that the PC has failed. You can see that there are many.

  FIG. 6 is a block diagram of the document processing apparatus of this embodiment. In the present embodiment, this apparatus includes a central processing unit CPU 601, a main memory 602, a display device 603, an input device 604, and a storage device 610. The storage device 610 includes an OS (operating system) 611, document data 612, public restriction word data 613, user information data 614, a user authentication program 615, a document search program 616, and a public document generation program 617. A document display program 618, word information data 619, a word information data generation program 620, and a word information data display program 621 are stored. The configuration different from the first embodiment is that word information data 619, a word information data generation program 620, and a word information data display program 621 are added to the present embodiment.

  The word information data 619 stores a set of words extracted from the document data, and the word information data 619 is generated by the word information data generation program 620. In addition, the word information data display program 621 displays the word information data 619 on the display device 603, and when a search request is received for the set of displayed words, the document search program 616 selects a document that matches the search condition. The document is identified and displayed on the display device 603.

  Next, the processing flow of the present embodiment will be described with reference to the flowchart of FIG. First, document data stored in a storage device such as a hard disk or a memory is read (S702), the document data is divided into words using morphological analysis, and word and part-of-speech information is acquired (S703). Next, using the word and part-of-speech information, a pair of words having a dependency relationship is acquired, and word information is generated by counting the number of word pairs (S704). For example, from the document “I want to repair my computer because it broke down”, “Computer (noun), ga (particle), breakdown (noun), shi (verb), ta (auxiliary verb), so (particle), repair (noun) ), Shi (verb), tai (auxiliary verb) "are obtained, and using the part of speech information of the word, word pairs" computer-failure "," repair-do "are extracted, and the number of word pairs is counted. Word information data as shown in FIG. Note that word pairs that have a dependency relationship are used here. However, even if they are not word pairs, it is possible to grasp the characteristics of a document such as a specified part-of-speech word or a set of words extracted under certain conditions. Any method is acceptable. Note that if the word information is generated in advance, the steps S702 to S704 are not necessary.

  Next, the access level of the user is specified by the user authentication means (S705), the open restriction word included in the word information is specified, and the type and the masking character string are set so that the same word included in different word pairs does not become the same masking character string. A masking character string of the disclosure restricted word composed of a set of randomly generated additional character strings is generated, and the word information to be disclosed is masked and disclosed by the masking character string (S706). In addition, the user authentication and the method for specifying the public restriction word are as described in the first embodiment. In addition, as described in the first embodiment, the masking character string may use color information or the like instead of the character string if the type is known. In addition, the additional character string is not generated randomly, but the alphabet from A to Z is used in order, for example, by using a method of extracting and using character strings in a predetermined order from a predetermined character string set. It doesn't matter.

Next, the word information to be displayed is narrowed down based on designated conditions such as the word pair with the highest frequency N rank, and the word information is displayed on a display device such as a display as in the example of FIG. 9 (S707).
Next, a document search request is received from the user with a word pair, a document including the word pair is extracted from the document data, and the document is displayed on the display device (S709). Here, when a word pair including a public restriction word is specified, the screen shown in FIG. 10 is displayed on the display device, indicating that the search cannot be performed because the public restriction word is included, and is included in the word pair. A confirmation screen asking whether or not to search with a word other than the restricted word to be displayed is displayed to allow the user to confirm whether or not the search is executed. For example, if there is a word pair “[Employee A] -Unfriendly” and a search request is received with this word pair, the word “[Employee A]” is a public restriction word, so the word “Unfriendly” A confirmation screen for whether to search is displayed. When “Yes” is selected, a document including the word “unfriendly” is searched, and when “No” is selected, the search is not executed. Therefore, even for a word pair including a public restriction word, it is possible to search for a word having no problem with a simple operation as a search condition. If there is no document search request, step S709 is skipped.

  According to the above embodiment, even if a word indicated by a masking character string included in word information such as a word pair is specified, a different masking character string is used for the same public restriction word included in another word pair. This has the effect of not affecting other word pairs. In addition, by restricting the search with word pairs including the restriction word, it is possible to prevent the restriction of the restriction word masked with the masking character string by browsing the text. In addition, there is an effect of facilitating the search for word pairs including masking character strings.

Next, a third embodiment of the present invention will be described with reference to the drawings.
The present embodiment is a document processing apparatus according to the second embodiment, in which a word pair including a public restriction word can be searched for a text by expanding the public restriction word into a common word according to the access right. FIG. 11 shows a public restriction word table including a public restriction word, a public restriction word access level, a public restriction word type, a superordinate concept, and a superordinate concept access level. Extend like a table. In the example of FIG. 11, one superordinate concept is set for each public restriction word. However, when a plurality of superordinate concepts are prepared according to the access level, the table is further expanded to correspond to one public restriction word. A plurality of superordinate concepts may be set.

  Next, the process flow of the present embodiment will be described. In the present embodiment, in the word information display screen described in the second embodiment, when a search request is received by a word pair and the word pair includes a public restriction word, the public restriction word database is referred to Referencing the concept access level, and determining whether or not the search is possible in the superordinate concept. If a search is possible, the fact that a high-level concept can be searched is displayed on the screen shown in FIG. 12, and a confirmation as to whether or not to search is requested. And the document is searched with all of the words and words other than the open restriction word included in the document information.

  For example, if the word pair “Yamada-san-unkind” is masked as “[Employee A] -unkind” and a user with access authority 2 issues a search request with this word pair, Since the concept is “Department A employee” and can be viewed by users with access authority 2, “Sato-san” with the same superordinate concept “Department A employee” is extracted from the public restricted word database, and “(Yamada-san or Sato-san) 3) Search the document using “unfriendly” as a search key. Since these publication restriction words included in the retrieved document are the same word as in the document publication apparatus described in the first embodiment, different masking character strings are used when the documents are different. The word indicated by the masking character string of each retrieved document is not specified.

  According to the present embodiment, even a word pair including a public restriction word can be expanded and searched for a word belonging to a higher concept according to the access authority. Therefore, the word pair including the public restriction word is used. It is possible to obtain a search result that is close to the case where the search that was made is secure and the public restriction word is used.

  This method masks the same word with different character strings and has little effect even if the word indicated by the masking character string is specified from some information. Applicable.

1 is a diagram illustrating a configuration diagram of a document disclosure apparatus (first embodiment). FIG. FIG. 10 is an explanatory diagram illustrating a processing flow of the document processing apparatus (first embodiment). (Example 1) which was the figure which showed the example of user information data. FIG. 6 is a diagram illustrating an example of document data (Example 1). (Example 1) which was the figure which showed the example of the publication | transmission restriction | limiting word data. FIG. 10 is a diagram illustrating a configuration diagram of a document processing apparatus (second embodiment). FIG. 10 is a diagram illustrating a processing flow of a document processing apparatus (second embodiment). (Example 2) which was the figure which showed the example of word information. It is the figure which showed the example of the masked word information (Example 2). (Example 2) which was the figure which showed the example of the confirmation screen of a document search. (Example 3) which was the figure which showed the example of the expansion | deployment restriction | limiting word data expanded. (Example 3) which was the figure which showed the example of the confirmation screen of a document search.

Explanation of symbols

  101: CPU, 102: main memory, 103: display device, 104: input device, 110: storage device, 111: OS, 112: document data, 113: disclosure restricted word data, 114: user information data, 115: user authentication Program: 116: Document search program, 117: Public document generation program, 118: Document display program.

Claims (9)

Obtaining user access rights;
Identifying a restriction word for public disclosure included in the public document based on the access authority;
Generating a masking character string composed of a character string indicating the type of the restricted word and a character string generated randomly or in a predetermined order;
A document publishing method comprising the step of displaying the restricted word contained in the published document by replacing it with the masking character string.   As the masking character string, color information for distinguishing a word is used for a character string indicating the type of the restricted word, or the kind of the restricted word is distinguished for a character string for distinguishing the word. The document publishing method according to claim 1, wherein the color information is used. A document publishing device having a storage device for storing a public document and publishing restricted word data, a calculation unit, and a display device,
The calculation unit specifies a public restriction word included in a public document stored in the storage device using the public restriction word data, and distinguishes a character string indicating the type of the public restriction word from a word in the type. Generating a masking character string that is combined with the character string to be used, and performing a replacement process of replacing the public restriction word included in the public document with the masking character string,
The display device displays a public document subjected to the replacement processing.   4. The document publishing apparatus according to claim 3, wherein color information is used as a character string for distinguishing words in the type. In a computing unit of a document publishing apparatus having a storage device for storing a public document and publication restricted word data, a computing unit, and a display device,
Reading the published document from the storage device;
Identifying a public restriction word included in the read public document using the public restriction word data;
Generating a masking character string that combines a character string indicating the type of the restricted word and a character string for distinguishing words in the type;
Performing the step of substituting the masking character string with the public restriction word included in the public document,
A document publishing program that causes the display device to display the replaced public document.   6. The document publishing program according to claim 5, wherein color information is used as a character string for distinguishing words in the type. Dividing document data into morphemes and dividing them into words;
Extracting one or more word sets from the divided words, counting the extracted word sets, and generating word information;
Identifying a public restriction word included in the words constituting the word set;
Generating a masking character string composed of a character string indicating the type of the restricted word and a character string generated randomly or in a predetermined order;
And a step of displaying the restricted word contained in the word information by replacing it with the masking character string.   When a search request is received for a set of words, and the public restriction word is included in the set of words, it is indicated that the search cannot be performed, the searchable conditions are presented, and whether or not the search is to be executed is selected. The document processing apparatus according to claim 7, wherein:   When a search request is received for a set of words, and when a public restriction word is included in the set of words and search is possible using a higher concept of the public restriction word, the fact that the higher concept can be searched is displayed, and search is executed The document processing apparatus according to claim 3, wherein whether or not is necessary is selected.

Images