JP2011170509A - Device, system, method and program for learning degree of secrecy - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a secrecy degree learning device which can reduce storage device resources and can reduce cost and effort required for introduction. <P>SOLUTION: The secrecy degree learning device includes: a management object storage 4 in which at least document management information comprising an attribute value which characterizes a document file is stored; a management rule database 5 which stores a management rule which makes a criterion of a management degree of the document file; and a management strictness determination part 3 which determines the management strictness of a document file specified by document storing information according to the management rule. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a confidentiality learning device, a confidentiality learning system, a confidentiality learning method, and a confidentiality learning program for evaluating the confidentiality of a document to be managed.

  In recent years, leakage of confidential information such as personal information has a great influence on corporate value, and information security management is being recognized as an important issue in corporate management. Such information security management requires sensitivity learning. Therefore, Patent Literature 1 and Patent Literature 2 propose the following confidentiality learning system.

  The confidentiality learning system disclosed in Patent Document 1 will be described with reference to the block diagram shown in FIG. This confidentiality learning system includes a confidential document database (DB) 118, a confidential document management server 119, a display 120 (120a, 120b,..., 120n), a client 121 (121a, 121b,. A confidential information management device 123, a proxy server 130, a gateway server 131, a facsimile server 132, a scan station 133, a portable information device 134, a network 135, and external devices 136 (136a, 136b,..., 136m) are provided.

  Also, the confidential information management device 123 includes a frequent expression extraction unit 124, a confidential information feature extraction unit 125, a confidential information determination unit 126, a hash table database (DB) 127, a determination log file 128, and a frequent expression table database (DB) 129. I have.

  In the confidentiality learning system, confidentiality learning processing that generates confidential information determination information stored in the hash table DB 127 and confidentiality evaluation processing that determines whether or not the evaluation target document includes confidential information. The processes are executed as independent processes.

  That is, in the confidentiality learning process, the confidential information management device 123 acquires the confidential document stored in the confidential document DB 118 via the confidential document management server 119 or the network 135. Then, the frequent expression extraction unit 124 of the confidential information management device 123 extracts a frequent expression from the acquired confidential document, and stores the extraction result in the frequent expression table DB 129.

  Thereafter, the confidential information feature extraction unit 125 generates information for determining confidential information based on the acquired confidential document and the information stored in the frequent expression table DB 129 and stores the information in the hash table DB 127.

  On the other hand, in the confidentiality evaluation process, the confidential information determination unit 126 obtains an evaluation target document obtained via the network 135, and the evaluation target document is classified based on the confidential information determination information stored in the hash table DB 127. It is determined whether or not information is included. Thereby, the confidentiality of the evaluation object document can be determined.

  Further, as shown in FIG. 19, the confidentiality evaluation system of Patent Document 2 includes a document reference unit 137, a region division unit 138, a feature element detection unit 139, a region-specific dictionary reference unit 140, a feature definition dictionary storage unit 141, a correlation. And a confidential information classification unit 143, a result output unit 144, and a document storage unit 145.

  The confidentiality evaluation system executes the confidentiality evaluation process, but does not perform the confidentiality learning process. In the confidentiality evaluation process, the document reference unit 137 reads the document stored in the document storage unit 145. Then, the area dividing unit 138 divides the read document into partial areas such as a header, a body, and a footer.

  The feature element detecting means 139 refers to the feature definition dictionary corresponding to the partial area, extracts the characteristic element from the partial area, and designates a candidate of a confidential information category that can be classified into the partial area.

  Correlation evaluation means 142 quantitatively evaluates the arrangement state of the feature elements corresponding to this category for each candidate confidential information category, and determines to which confidential information category the partial area is classified. .

  The confidential information classification unit 143 determines which confidential information category the document is classified on the basis of the confidential information category into which each partial area is classified and the importance of each confidential information category. Determine the importance. This determines the importance of the document.

Republished patent 2007-105273 JP 2006-209649 A

  However, in the above-described confidentiality learning system and confidentiality evaluation system, there are problems that the system introduction cost is high and a great deal of labor is required for system introduction.

  This is because the confidential document DB 118 in Patent Document 1 and the feature definition dictionary storage means 141 in Patent Document 2 must be prepared for each system installation case and each type of document to be managed.

  Further, the information stored in the confidential document DB 118 in Patent Document 1 and the feature definition dictionary storage unit 141 in Patent Document 2 needs to be maintained such as addition or update in accordance with the change in confidentiality of the document to be managed. . Such maintenance requires a large amount of money and a great deal of labor. In addition, there is a problem that the consumption of storage device resources for storing such information increases.

  In particular, in the confidentiality learning system of Patent Document 1, since the confidential document for learning is separately stored in the confidential document DB 118, the storage device resource is consumed significantly.

  Accordingly, the main object of the present invention is to reduce the cost and labor required after the introduction of the system and reduce the consumption of storage device resources, the sensitivity learning system, the sensitivity learning system, the sensitivity learning method, and the sensitivity learning. Is to provide a program.

  In order to solve the above-mentioned problem, an invention relating to a confidentiality learning device includes a management target storage in which at least document storage information including attribute values characterizing a document file is stored, and a management that is a criterion for determining the management degree of the document file. A management rule database for storing rules, and a management severity determination unit that determines the management severity of the document file specified by the document storage information in accordance with the management rules.

  The confidentiality learning system includes the confidentiality learning device and a user terminal that causes the confidentiality learning device to specify the evaluation document file and to evaluate the confidentiality. To do.

  Further, the invention according to the confidentiality learning method is specified by the document storage information based on document storage information including attribute values that characterize the document file and a management rule that is a criterion for determining the management degree of the document file. A management severity determination procedure for determining the management severity of the document file is included.

  Further, the invention relating to the confidentiality learning program is specified by the document storage information based on the document storage information including attribute values characterizing the document file and the management rule that determines the management degree of the document file. A management severity determination step for determining the management severity of the document file is included.

  According to the present invention, since confidentiality learning is determined based on the management rules stored in the management rule DB, only the management rules stored in the management rule DB are prepared when introducing the confidentiality learning system. You can do it. Therefore, it is possible to reduce the storage device resources and reduce the cost and labor required for introduction.

1 is a block diagram of a confidentiality learning apparatus according to a first embodiment of the present invention. It is a block diagram of the confidentiality learning system concerning the 2nd Embodiment of this invention. It is a figure which shows the table structure of the document storage information table concerning 2nd Embodiment. It is a figure which shows the table structure of the management rule table concerning 2nd Embodiment. It is a figure which shows the table structure of the management severity table concerning 2nd Embodiment. It is a flowchart of the learning process concerning 2nd Embodiment. It is a flowchart of the severeness determination process concerning 2nd Embodiment. It is a flowchart of the evaluation process concerning 2nd Embodiment. It is a block diagram of the confidentiality learning system concerning the 3rd Embodiment of this invention. It is a flowchart of the all document confidentiality evaluation process concerning 3rd Embodiment. It is a block diagram of a confidentiality learning system concerning a 4th embodiment of the present invention. It is a figure which shows the table structure of the document information table concerning 4th Embodiment. It is a figure which shows the table structure of the transposed index table concerning 4th Embodiment. It is a flowchart of the learning process concerning 4th Embodiment. It is a flowchart of the management degree learning process in the learning process concerning 4th Embodiment. It is a flowchart of the evaluation process concerning 4th Embodiment. It is a flowchart of the management degree evaluation process concerning 4th Embodiment. It is a block diagram of a confidentiality learning system concerning related technology. It is a block diagram of a confidentiality evaluation system concerning related technology.

<First Embodiment>
A first embodiment of the present invention will be described. FIG. 1 is a block diagram of the confidentiality learning apparatus according to the present embodiment. The confidentiality learning device 2 includes a management target storage 4, a management rule database (DB) 5, and a management severity determination unit 3. The management target storage 4 stores at least document storage information including attribute values that characterize document files. The management rule database 5 stores management rules that serve as criteria for determining the degree of management of document files. The management severity determination unit 3 determines the management severity of the document file specified by the document storage information according to the management rule.

As a result, in order to make it possible to determine the management severity of the document file, the user only needs to prepare the management rules of the management rule database 5, so that the introduction cost and the introduction labor of the confidentiality learning device can be reduced. become.
<Second Embodiment>
Next, a second embodiment of the present invention will be described. FIG. 2 is a block diagram of the confidentiality learning system 9B according to the present embodiment. The confidentiality learning system 9B includes a confidentiality learning device 10B and a user terminal 11. Further, the confidentiality learning device 10B includes an input / output unit 12a, a management target storage 13 that is a repository for storing management target files, a management severity determination unit 14, a management rule DB 15, a management severity DB 16, and a sensitivity evaluation unit 17. It has.

  The management target storage 13 stores a document file and a document storage information table. This document file means the document itself. Therefore, in the following description, in order to avoid confusion with other terms, it is described as a document file.

  FIG. 3 is a diagram showing a table configuration of the document storage information table 50 stored in the management target storage 13. The document storage information table 50 includes a plurality of document storage information 55 including a plurality of attribute information related to document files. One document storage information 55 includes file path information 51 that uniquely indicates the storage location of one document file, authority setting information 52 that indicates access authority to the document file, and encryption that indicates whether or not the document file is encrypted. And attribute information such as archiving information 53 and custodian information 54 indicating information such as the owner of the document file. In FIG. 3, a blank 56 is shown, which indicates that attribute information can be added. For example, there are cases where it is desired to limit the persons who can view the document file. In such a case, new attribute information for “browser restriction” is provided in the blank 56.

  For example, the uppermost document storage information 55 in FIG. 3 is “a document file in which the file path information 51 is stored at the position of [C: ¥ file.text] is specified by the authority setting information 52 of [777]. It is managed by an authorized person and is not encrypted because the encryption information 53 is [None], and the custodian of this document file is [Yamada] indicated by the custodian information 54.

  The file path information 51 only needs to be able to uniquely identify the storage location of the document file. Therefore, the file path information 51 can use a file path depending on the OS (Operating System), a URI (Uniform Resource Identifier), a numerical value, a character string, and the like.

  Examples of the authority setting information 52 include a combination of a user and an access authority, a notation using a group that is a set of users, and the like. Examples of the encryption information 53 include information such as binary information indicating the presence / absence of encryption, encryption strength, and encryption type. As the custodian information 54, a custodian or manager of a document file can be exemplified.

  The management severity determination unit 14 determines the management severity of the document file stored in the management target storage 13 based on the management rule stored in the management rule DB 15. The management severity is information indicating the security level of the document file. The management rule is a criterion for determining the severity of management for a document file.

  FIG. 4 is a diagram illustrating a table configuration of the management rule table 60. This management rule table 60 includes a plurality of management rules 64. One management rule 64 is assigned to rule ID information 61 for identifying (specifying) the management rule 64, rule information 62 forming a rule entity of the management rule 64, and a document file that matches the rule information 62. Severity information 63 indicating the severity is included. The severity information 63 is information that expresses the confidentiality of the document file to which the management rule 64 is applied by a numerical value.

For example, the management rule 64 at the top of FIG. 4 indicates that “the rule whose rule ID information 61 is [1] sets the severity information 63 of the document file to [0. 9] ”. Note that the rule ID information 61 is information that uniquely identifies the management rule 64, and numerical values, character strings, URIs, and the like can be used.
Also,
Then, the management severity determination unit 14 acquires a document file from the management target storage 13 via a local file system or a network (not shown). Further, the management severity determination unit 14 acquires document storage information 55 for the document file acquired from the document storage information table 50 stored in the management target storage 13. Then, the management severity of this document file is evaluated by comparing the document storage information 55 of the acquired document file with the management rule 64 stored in the management rule DB 15.

  The management severity DB 16 stores the severity determined by the management severity determination unit 14 as the severity information 63. FIG. 5 is a diagram showing a table configuration of the management severity table 70. This management severity table 70 includes a plurality of management severity information 73. The management severity information 73 includes file path information 71 and severity information 72. The file path information 71 corresponds to the file path information 51 shown in FIG. 3, and the severity information 72 corresponds to the severity information 63 shown in FIG. However, as described later, since the severity information 72 may be an average value of a plurality of severity information 63, the severity information 72 and the severity information 63 are not uniquely the same.

  The input / output unit 12a receives a request for evaluating the confidentiality of the document file designated by the user via the user terminal 11. Hereinafter, the document file requested for evaluation is referred to as an evaluation document file. Then, the input / output unit 12a causes the confidentiality evaluation unit 17 to evaluate the confidentiality of the received evaluation document file. The confidentiality evaluation unit 17 evaluates the confidentiality and outputs the evaluation result to the input / output unit 12a. The input / output unit 12 a outputs the confidentiality evaluation result to the user terminal 11.

  The confidentiality evaluation unit 17 includes a confidentiality learning unit 18, a confidentiality learning result DB 19, and a confidentiality evaluation unit 20. Then, the confidentiality learning unit 18 learns the confidentiality of the document file using the management severity information 73 stored in the management severity DB 16 and the document storage information 55 stored in the management target storage 13. This confidentiality is stored in the confidentiality learning result DB 19.

  When the confidentiality evaluation unit 20 receives a request for evaluating the confidentiality of the evaluation document file from the input / output unit 12a, the confidentiality evaluation unit 20 evaluates the confidentiality of the evaluation document file based on the confidentiality stored in the confidentiality learning result DB 19. The evaluation result is output to the input / output unit 12a.

  Next, processing in the management severity determination unit 14 and the confidentiality learning unit 18 (hereinafter referred to as learning processing) and evaluation in the confidentiality evaluation unit 20 (hereinafter referred to as evaluation processing) will be described in detail.

  First, the learning process will be described. FIG. 6 is a flowchart of the learning process. As described above, the learning process mainly includes the severity determination process (step Sa1) and the confidentiality learning process (step Sa2).

  In the severity determination processing, the severity of the document file stored in the management target storage 13 is determined by the management severity determination unit 14, and the determination result is registered in the management severity DB 16 as management severity information 73. It is processing.

  The confidentiality learning process uses the confidentiality learning unit 18 to learn the confidentiality of the document file using the document file stored in the management target storage 13 and the management severity information 73 stored in the management severity DB 16. It is processing to do.

  FIG. 7 is a flowchart of the severity determination process. The management severity determination unit 14 acquires the document file stored in the management target storage 13, and acquires the document storage information 55 of the document file from the document storage information table 50 (step Sb1).

  Next, the management severity determination unit 14 acquires the management rule 64 from the management rule DB 15, and compares the management rule 64 with the document storage information 55 (step Sb2). If the document storage information 55 and the management rule 64 match, the process proceeds to step Sb3, and if they do not match, the process proceeds to step Sb4.

  For example, it is assumed that the document file to be determined is characterized by the document storage information 55 on the second line from the top in the document storage information table 50 shown in FIG. In this document storage information 55, the file path information 51 is [file: /// c: \ doc1. doc], the authority setting information 52 is [700], the encryption information 53 is [Yes], and the custodian information 54 is [mail: //Yamada@...jp]. When it is determined that the plurality of management rules 64 shown in FIG. 4 match the document storage information 55, the [encryption] rule information 62 matches. The severity of the document file that matches the management rule 64 is set to [0.9] from the severity information 63.

  When the document storage information 55 matches the management rule 64, the management severity determination unit 14 temporarily stores the severity information 63 of the matched management rule 64 (step Sb3).

  The management rule table 60 includes a plurality of management rules 64. Therefore, the management severity determination unit 14 determines whether the severity information 63 has been compared with all the management rules 64 (step Sb4).

  When the severity level information 63 is compared with all the management rules 64, the management severity level determination unit 14 determines the severity level of the document file to be determined based on the severity level information 63 of the matched management rule 64. The degree information 72 is calculated (step Sb5). Since a plurality of management rules 64 exist, a plurality of management rules 64 that match the document storage information 55 may be detected. Therefore, the severity information 72 of the document file can be an average value, median value, or maximum value of the severity information 63 in each management rule 64.

  The calculated severity of the document file and the file path information 71 included in the document storage information 55 are registered in the management severity DB 16 as the management severity information 73 (step Sb6).

  It is determined whether or not the above severeness determination process has been performed for all the document files (step Sb7). If the severity determination processing has not been performed for all the document files, the process returns to step Sb1, and if it has been performed for all the document files, the processing ends.

  The confidentiality learning unit 18 learns confidentiality using the management severity information 73 and the document storage information 55 determined in this way. The confidentiality learning result is stored in the confidentiality learning result DB 19 as the confidentiality learning result. As a specific learning procedure of the confidentiality in the confidentiality learning unit 18, a known confidentiality learning method or the like can be used.

  Next, the evaluation process will be described. FIG. 8 is a flowchart of the evaluation process. First, the confidentiality evaluation unit 20 in the confidentiality evaluation unit 17 waits for a confidentiality evaluation request via the input / output unit 12a (steps Sc1 and Sc2). In the confidentiality evaluation request, an evaluation document file to be evaluated is designated by designating a document file name or file path information or by directly inputting a document file.

  When receiving the evaluation document file, the confidentiality evaluation unit 20 evaluates the confidentiality (step Sc3). The evaluation result of the confidentiality is sent to the input / output unit 12a.

  The input / output unit 12a outputs the confidentiality evaluation result of the evaluation document file received from the confidentiality evaluation unit 20 to the user terminal 11 (step Sc4).

  As described above, the confidentiality learning is performed based on the management severity information shown in FIG. Since this management severity information is determined based on the management rules stored in the management rule DB, it is only necessary to prepare the management rules stored in the management rule DB when introducing the confidentiality learning system. become.

  The management severity information includes two pieces of information, file path information and severity information. Therefore, compared with the confidential document DB and the feature definition dictionary described in the related art, the amount of information that the confidentiality learning system according to the present invention has to prepare in advance is very small.

  Therefore, the introduction cost of the confidentiality learning system according to the present invention is reduced, and the labor required for the introduction is greatly reduced.

  In addition, since many management rules stored in the management rule DB are rules with high universality and versatility, depending on the purpose of introducing the system as described in the related technology and the characteristics of the document file to be managed. It is not necessary to prepare and adjust the confidential document DB and the feature definition dictionary. Therefore, also from this viewpoint, the introduction cost of the confidentiality learning system according to the present invention can be reduced, and the labor required for introduction can be reduced.

  Furthermore, since confidentiality learning can be performed only with information stored in the document storage information storage or the management rule DB, it is possible to reduce maintenance costs and reduce maintenance labor, and to reduce storage device resources.

In particular, the management rule DB is information with a low frequency of information change, such as a document encryption rule in the storage to be managed, so that the labor and maintenance cost associated with the update can be greatly reduced. Further, since the management rule DB is smaller than the confidential document DB and the feature definition dictionary in the related technology, the storage device resources can be greatly reduced.
<Third Embodiment>
Next, a third embodiment of the present invention will be described. In addition, about the same structure as 2nd Embodiment, description is abbreviate | omitted suitably using the same code | symbol. FIG. 9 is a block diagram of a confidentiality learning system 9C according to the present embodiment.

  The confidentiality learning device 10C in the confidentiality learning system 9C is different from the confidentiality learning device 10B shown in FIG. 2 in that an all document evaluation execution unit 21 is additionally provided.

  The input / output unit 12b receives designation of the evaluation document file from the user terminal 11, and instructs the confidentiality evaluation unit 20 to perform confidentiality evaluation on the evaluation document file. The input / output unit 12 b receives the result of the confidentiality evaluation from the confidentiality evaluation unit 20 and outputs the result to the user terminal 11.

  Further, the input / output unit 12 b accepts an execution instruction for confidentiality evaluation for all document files included in the management target storage 13 from the user terminal 11. Upon receiving an instruction to execute confidentiality evaluation for all document files, the input / output unit 12b instructs the all document evaluation execution unit 21 to evaluate confidentiality for all document files. Then, the input / output unit 12b outputs the result of the confidentiality evaluation to the user terminal 11.

  Upon receiving an instruction to execute confidentiality evaluation for all document files from the input / output unit 12b, the all document evaluation execution unit 21 causes the management severity determination unit 14 to execute a severity determination process. Further, the all document evaluation execution unit 21 instructs the confidentiality evaluation unit 20 to perform all document confidentiality evaluation processing using the result of the severity determination processing. As a result, the confidentiality of all document files is evaluated. Therefore, the all document evaluation execution unit 21 rearranges the evaluation results by the confidentiality evaluation unit 20 in the order of confidentiality and outputs them to the input / output unit 12b.

  Such a total document density evaluation process will be described with reference to the flowchart of FIG. First, when the management severity determination unit 14 receives an execution instruction for the severity determination processing from the all document evaluation execution unit 21, the management severity determination unit 14 executes the severity determination processing (step Sd1).

  Next, the all document evaluation execution unit 21 reads a document file from the management target storage 13 (step Sd2), and sends the read document file to the confidentiality evaluation unit 20.

  The confidentiality evaluation unit 20 executes the evaluation process as shown in FIG. 8 on the document file received from the all document evaluation execution unit 21 (step Sd3). The evaluation result is output to the all document evaluation execution unit 21.

  Then, it is determined whether or not such evaluation processing has been executed for all document files (step Sd4). If the evaluation process has been executed for all document files, the process proceeds to step Sd4. If there is a document file that has not been executed yet, the process returns to step Sd2.

  The all document evaluation execution unit 21 rearranges the list of sets of the acquired document file and confidentiality of the document file in descending order of confidentiality and outputs the sorted list to the input / output unit 12b (step Sd5).

  The input / output unit 12b outputs a list of pairs of the obtained document file and confidentiality to the user terminal 11 (step Sd6).

  As described above, the all document evaluation execution unit 21 automatically evaluates the confidentiality of all the document files as a series of processes, so that the operation cost and the operation labor can be reduced.

In addition, since all the document files stored in the management target storage can be easily evaluated by the all document evaluation execution unit 21, the operation cost and the operation labor can be reduced.
<Fourth Embodiment>
Next, a fourth embodiment of the present invention will be described. In addition, about the same structure as 2nd Embodiment, description is abbreviate | omitted suitably using the same code | symbol. FIG. 11 is a block diagram of a confidentiality learning system 9D according to the present embodiment.

  The confidentiality learning device 10D of the confidentiality learning system 9D is additionally provided with a management degree evaluation unit 22 and a required countermeasure degree calculation unit 27 with respect to the confidentiality learning apparatus 10B shown in FIG.

  The management degree evaluation unit 22 learns the management degree of the document file based on the management severity information 73 stored in the management severity DB 16 and the document storage information 55 stored in the management target storage 13. Evaluate the management level of the specified evaluation document file. Such a management level evaluation unit 22 includes an index generation unit 23, an index DB 24, a search unit 25, and a management level evaluation unit 26.

  The index DB 24 is a database that stores index information and document information generated by the index generation unit 23, and includes an inverted index table 80 and a document information table 90.

  The document information table 90 includes a plurality of document information 97 as shown in FIG. One piece of document information 97 includes document ID information 91 for identifying a document file, file path information 92 that uniquely indicates a storage location of the document file, authority setting information 93 that indicates an access authority for the document file, and the document. It includes encryption information 94 indicating whether or not the file is encrypted, and severity information 95 indicating the severity of management for the document file. In the document information table 90 shown in FIG. 12, a blank space 96 is provided. This blank space 96 indicates that this information is used when other information different from the information described above is added. . That is, the number of attribute items of the document information 97 is provided so that it can be appropriately increased or decreased.

  As shown in FIG. 13, the inverted index table 80 stores a plurality of inverted index information 83. One transposed index information 83 includes item information 81 such as a word included in the document file, and document ID group information 82 including document ID information 91 of a plurality of document files including the item information 81. For example, in the inverted index information 83 whose item information 81 is [Tokyo], the document ID group information indicating that the document files whose document ID information 91 is [1], [2], [5], [. I have 82.

  The index generation unit 23 acquires a document file from the management target storage 13 via a local file system or a network (not shown). And the index production | generation part 23 extracts character string information from the acquired document file, and extracts an item from this character string information. Since the document ID 91 is assigned to the document file including the extracted item, the document ID 91 is acquired. The collected document ID 91 forms document ID group information 82, and the item forms item information 81. Therefore, the index generation unit 23 adds the item information 81 and the document ID group information 82 as the inverted index information 83 to the inverted index table 80 of the index DB 24.

  Further, the index generation unit 23 acquires the document storage information 55 as shown in FIG. 3 from the management target storage 13 and also acquires the management severity information 73 as shown in FIG. 5 from the management severity DB 16. Then, the index generation unit 23 extracts file path information 51, authority setting information 52, and encryption information 53 from the document storage information 55, and these are extracted as file path information 92 corresponding to the document ID information 91, authority setting information 93, The encryption information 94 is assumed. Further, the index generation unit 23 extracts the severity information 72 from the management severity information 73 and sets it as the severity information 95 corresponding to the document ID information 91. Thereby, the document ID information 91, the file path information 92, the authority setting information 93, the encryption information 94, and the severity information 95 are acquired. Then, the index generation unit 23 adds these as document information 97 to the document information table 90 of the index DB 24.

  The search unit 25 creates a query based on the character string information of the evaluation document file for which an evaluation request has been made, and searches the transposed index table 80 of the index DB 24. Then, a document file having contents similar to the contents of the evaluation document file is acquired.

  The management level evaluation unit 26 evaluates the management level of the evaluation document file. That is, the management level evaluation unit 26 acquires a document file group similar to the evaluation document file searched by the search unit 25 and document information of the document file group. Then, the management degree evaluation unit 26 performs statistical processing for quantifying the management status of the similar document file group with respect to the management severity of the similar document file group, and sets the management degree as an evaluation result.

  The required degree-of-hands calculation unit 27 quantifies the necessity of changing the storage state for the evaluation document file input by the input / output unit 12a, and calculates this as the required degree of action.

  The required degree of handling is calculated based on the confidentiality of the evaluation document file evaluated by the confidentiality evaluation unit 20 and the management degree of the evaluation document file evaluated by the management degree evaluation unit 26. The required response level calculated by the required response level calculation unit 27 is output to the input / output unit 12a.

  Next, a processing procedure of such a confidentiality learning system 9D will be described. The confidentiality learning system 9D uses strictness determination processing, learning processing, and evaluation processing as main processing procedures. The severity determination process is the same as that in the second embodiment, but the learning process and the evaluation process are different, so the learning process and the evaluation process will be described in detail.

  FIG. 14 is a flowchart of the learning process. In the learning process, after the severity determination process (step Sa1) is performed, the confidentiality learning process (step Sa2) and the management degree learning process (step Se1) are performed. 14 to 16 are diagrams according to a diagram defined by UML (Unified Modeling Language), and the black painting step in the figure indicates [branching point PS] or [confluence point PE]. The execution order of a series of sequences between the branch point PS and the merge point PE is not defined. For example, in FIG. 14, the confidentiality learning process and the management degree learning process are included between the branch point PS and the confluence point PE, but the confidentiality learning process and the management degree learning process may be performed simultaneously. One process may be executed prior to the other process.

  Since the confidentiality learning process (step Sa2) is described in FIG. 6, the management degree learning process (step Se1) will be described in detail below.

  This management degree learning process uses all the document files stored in the management target storage 13 by the index generation unit 23 and the management severity information 73 stored in the management severity DB 16 and the inverted index information 83 and This is a procedure for creating the document information 97 and registering it in the index DB 24.

  FIG. 15 is a flowchart of the management level learning process in the learning process. When the management degree learning process is started, three sequences (a process for creating / registering file path information 92 and the like in steps Sf1 to Sf4, and a process for creating and registering document information 97 in steps Sf5 to Sf6) are started from the branch point PS. , Processing for creating / registering the severity information 95 in steps Sf7 to Sf8).

  The index generation unit 23 acquires a document file from the management target storage 13 via a local file system or a network (Step Sf1). Then, the index generation unit 23 extracts the document ID information 91 and the character string included in the acquired document file (step Sf2), and extracts item information 81 from the character string (step Sf3). As the item information 81, a word or a character string by the N character index method (n-gram) can be applied. Thereafter, the index generation unit 23 creates the inverted index information 83 using the extracted item information 81 and registers it in the inverted index table 80 of the index DB 24 (step Sf4).

  Further, the index generation unit 23 acquires the document storage information 55 of the document file corresponding to the document ID information 91 from the management target storage 13 via the local file system or the network (step Sf5).

  Then, the index generation unit 23 registers the file path information 92, authority setting information 93, and encryption information 94 from the acquired document storage information 55 in the document information table 90 (step Sf6).

  Further, the index generation unit 23 acquires the management severity information 73 from the management severity DB 16 and extracts the severity information 72 of the corresponding document file from the management severity information 73 (step Sf7). Then, the index generation unit 23 registers the severity information 72 as the severity information 95 in the document information table 90 of the index DB 24 (step Sf8).

  When such creation / registration processing of the inverted index information 83, creation / registration processing of the document storage information 55, and creation / registration processing of the severity information 95 are executed, the information is stored in the management target storage 13 via the junction PE. It is determined whether or not the above-described processing has been performed on all the document files that have been processed (step Sf9). If the processing has not been completed for all the document files, the process returns to the branch point PS. If the processing has been completed, the management degree learning process ends.

  Next, the evaluation process will be described. FIG. 16 is a flowchart of this evaluation process. In the evaluation process, the acceptance of the evaluation document file is determined (steps Sc2 and Sc1). When the evaluation document file is received, the confidentiality evaluation process (step Sc3) and the management degree evaluation process (step Sg1) are performed. Thereafter, the confidentiality and management level of the evaluation document file are calculated by the required level calculation unit 27, and the required level of response of this evaluation document file is calculated (Sg2).

  Next, the management degree evaluation process will be described in detail. Note that the confidentiality evaluation process (step Sc3) has been described in the previous embodiment, and therefore redundant description is avoided. FIG. 17 is a flowchart of the management degree evaluation process.

  First, the retrieval unit 25 retrieves a document file similar to the evaluation document file from the transposed index table 80 of the index DB 24, and obtains a similar document file group (step Sh1).

  Next, the search unit 25 acquires the severity information 95 for each document file in the similar document file group from the document information 97 stored in the index DB 24 (step Sh2).

  Then, the management level evaluation unit 26 evaluates the management level of the evaluation document file, and calculates a variance value, a median value, and the like indicating the consistency between the management level and the severity information 95 of the similar document file group by statistical processing. (Step Sh3).

  As described above, the management degree obtained by quantifying the index of the management state of the document file evaluated by the management degree evaluation unit is acquired, and the sensitivity evaluated by the sensitivity evaluation unit is acquired. The degree of coping required is calculated from the density. Therefore, whether or not the user should improve the management state of the management target storage 13 can be determined only from the degree of necessity of the evaluation document file. In other words, a document file with a poor management level despite a high sensitivity is calculated with a high handling level. Therefore, it is possible to easily determine whether to improve the management state of the management target storage 13 and to manage the state.

  In addition, various processing procedures in the confidentiality learning system can be coded and programmed to be executable by a computer, and the program can be recorded on an information recording medium.

  The features described above are summarized as additional notes. APPENDIX 1 is a management target storage that stores at least document storage information including attribute values that characterize a document file, a management rule database that stores a management rule that serves as a criterion for determining the degree of management of the document file, and a document according to the management rule. A confidentiality learning apparatus comprising: a management severity determining unit that determines a management severity of a document file specified by storage information.

  Appendix 2 is the confidentiality learning apparatus according to Appendix 1, further comprising a confidentiality evaluation unit that learns the confidentiality of the document file based on the management severity and the document file by the management severity determination unit. Sensitivity learning device.

  Supplementary Note 3 is the confidentiality learning apparatus according to Supplementary Note 2, wherein the confidentiality evaluation unit includes a confidentiality learning unit that performs learning for confidentiality evaluation using the management severity as the confidentiality of the document file. Sensitivity learning device.

  Appendix 4 is the confidentiality learning device according to Appendix 2 or 3, wherein the confidentiality evaluation unit includes a confidentiality evaluation unit that evaluates the confidentiality of the designated evaluation document file based on the learning result of the confidentiality learning unit. A confidentiality learning apparatus comprising:

  Appendix 5 is the confidentiality learning apparatus according to Appendix 1, further comprising an all document evaluation execution unit that instructs the confidentiality evaluation unit to evaluate the confidentiality of all document files stored in the management target storage. Sensitivity learning device.

  Appendix 6 is the confidentiality learning device according to Appendix 3 or 4, further comprising a management degree evaluation unit that evaluates the management degree of the evaluation document file based on the management severity and the document storage information. apparatus.

  Supplementary Note 7 is a confidentiality learning device according to any one of Supplementary Notes 4 to 6, in which the degree of required action indicating the necessity of management of the evaluation document file by the user is calculated based on the confidentiality and the management degree. A confidentiality learning apparatus comprising a first-to-first-comparison calculating unit.

  The supplementary note 8 includes the confidentiality learning device according to any one of the supplementary notes 1 to 7, and a user terminal that causes the confidentiality learning device to designate an evaluation document file and to evaluate the confidentiality. Sensitive learning system characterized by that.

  Appendix 9 determines the management severity of the document file specified by the document storage information based on the document storage information including attribute values that characterize the document file and the management rule that serves as a criterion for determining the management level of the document file. A confidentiality learning method including a management severity determination procedure.

  Appendix 10 is the confidentiality learning method according to appendix 9, wherein the confidentiality learning method includes a learning procedure for learning for confidentiality evaluation using the management severity as the confidentiality of the document file.

  Appendix 11 is the confidentiality learning method according to appendix 10, wherein the confidentiality learning method includes a procedure for evaluating the confidentiality of the designated evaluation document file based on the confidentiality learning result.

  Appendix 12 is the confidentiality learning method according to appendix 9 or 10, wherein the confidentiality learning method includes a procedure for evaluating the management degree of the evaluation document file based on the management severity and the document storage information.

  The supplementary note 13 is characterized in that, in the confidentiality learning method according to the supplementary note 12, a procedure for calculating a required action level indicating the necessity of management of the evaluation document file by the user based on the confidentiality and the management degree is characterized. Sensitivity learning method to do.

  The supplementary note 14 determines the management severity of the document file specified by the document storage information based on the document storage information including attribute values that characterize the document file and the management rule that is a criterion for determining the management level of the document file. A confidentiality learning program including a management severity determination step.

  Supplementary Note 15 is a confidentiality learning program according to the supplementary note 14, including a learning step of performing learning for confidentiality evaluation using the management severity as the confidentiality of the document file.

  Supplementary Note 16 is a confidentiality learning program according to the supplementary note 15, including a step of evaluating the confidentiality of the designated evaluation document file based on the confidentiality learning result.

  Supplementary Note 17 is a confidentiality learning program according to any one of Supplementary Notes 14 to 16, and includes a step of evaluating the management degree of the evaluation document file based on the management severity and the document storage information. program.

  Supplementary Note 18 is characterized in that, in the confidentiality learning program according to Supplementary Note 17, a step of calculating a required action level indicating the necessity of management of the evaluation document file by the user based on the confidentiality and the management degree is characterized. A sensitive learning program.

9B-9D Sensitivity learning system 2, 10B-10D Sensitivity learning device 12a, 12b Input / output unit 4,13 Managed storage 3,14 Management severity determination unit 5,15 Management rule DB
16 Management severity DB
17 Sensitivity evaluation unit 18 Sensitivity learning unit 19 Sensitivity learning result DB
20 Sensitivity Evaluation Unit 21 All Document Evaluation Execution Unit 22 Management Level Evaluation Unit 23 Index Generation Unit 24 Index DB
25 Retrieval Unit 26 Management Level Evaluation Unit 27 Action Required Level Calculation Unit

Claims (10)

Managed storage for storing at least document storage information consisting of attribute values characterizing document files;
A management rule database that stores management rules that form criteria for determining the degree of management of the document file;
A confidentiality learning apparatus comprising: a management severity determining unit that determines a management severity of the document file specified by the document storage information according to the management rule. The confidentiality learning apparatus according to claim 1,
A confidentiality learning apparatus comprising: a confidentiality evaluation unit that learns the confidentiality of the document file based on the management severity and the document file by the management severity determination unit. The confidentiality learning apparatus according to claim 2,
The confidentiality evaluation unit includes a confidentiality evaluation unit that evaluates a confidentiality of a designated evaluation document file based on a learning result of the confidentiality learning unit. The confidentiality learning apparatus according to claim 3,
A confidentiality learning apparatus comprising: a management degree evaluation unit that evaluates a management degree of the evaluation document file based on the management severity and the document storage information. The confidentiality learning apparatus according to claim 4,
A confidentiality learning apparatus, comprising: a point-to-initial degree calculation unit that calculates a degree of necessity that indicates a necessity of management of the evaluation document file by a user based on the degree of confidentiality and the degree of management. The confidentiality learning device according to claim 3,
A confidentiality learning system comprising: a user terminal that causes the confidentiality learning device to specify the evaluation document file and perform confidentiality evaluation.   Management for determining the management severity of the document file specified by the document storage information based on document storage information comprising attribute values characterizing the document file and a management rule that is a criterion for determining the management level of the document file A confidentiality learning method characterized by including a severity determination procedure. The confidentiality learning method according to claim 7,
A confidentiality learning method, comprising: a learning procedure for performing learning for confidentiality evaluation using the management severity as confidentiality of the document file. The confidentiality learning method according to claim 8,
A confidentiality learning method comprising a step of evaluating the confidentiality of a designated evaluation document file based on the learning result of the confidentiality.   Management for determining the management severity of the document file specified by the document storage information based on document storage information comprising attribute values characterizing the document file and a management rule that is a criterion for determining the management level of the document file A confidentiality learning program including a severity determination step.

Images