JP3826151B2 - Information maintenance management apparatus and method for information sharing system - Google Patents

Description

  The present invention provides information such as knowledge, know-how possessed by individuals, documents created by individuals, computer software, copyrighted works such as paintings, and knowledge of fields owned by a specific group on a computer network. The present invention relates to an information sharing system that can be made public and shared via a computer network.

  In recent years, electronic bulletin boards on the computer network, Internet homepages, and the like have been used as means for sharing information on the computer network. As the network environment has been improved and the prices of information communication devices and storage devices have been reduced, individuals and organizations can easily send and receive information using these means.

  For example, many homepages have been established on the Internet, and a variety of information can be examined by using a search service on the Internet. In a local network, information transmission and reception within an organization can be performed using, for example, an electronic bulletin board on a personal computer.

  Under such an environment, information provision itself has been performed as a service on the Internet. For example, there is a place where a help desk, which is an inquiry window from a manufacturer's customer, discloses useful troubleshooting information to contract users via the Internet. In such a help desk, the quality and quantity of information itself provided is important because it differentiates it from other help desks.

  Conventional information sharing means have few functions for maintaining the quality of information. However, the quality of information deteriorates over time, such as the contents becoming incorrect. For example, the troubleshooting information at the help desk needs to be inspected due to changes such as improvement of product functions. The information sharing means described above is not provided with a function for supporting such inspection.

  In the end, in order to maintain the quality of information in a conventional information sharing system, all information accumulated by humans must be reviewed periodically and corrected as necessary. This is not a very practical means in the recent situation where information transmission is simplified and a large amount of information is accumulated.

  Therefore, with traditional information sharing, you have the choice between accumulating large amounts of low quality information that you do n’t know is right or useful, or narrowing down the amount of information to maintain quality, A large amount of information that can be guaranteed to be useful to users could not be disclosed.

  The present invention has been made in view of the above problems, by detecting what needs to be disclosed from accumulated non-public information, giving priority to it, and notifying the manager of shared information. An object of the present invention is to provide an information sharing system capable of efficiently reviewing information.

In order to achieve the above object, an information maintenance management device of an information sharing system of the present invention is connected to a computer network, and a public database for publishing accumulated information in response to a user's inquiry, and a public object Non-public databases that store non-public information and user queries to the public database, or information that refers to the private database as a query to the private database, and the corresponding information is In some cases, the effectiveness frequency is calculated by adding the score corresponding to the effectiveness ranking of the information, and this is compared with the effectiveness frequency of other information allocated in advance, and the effectiveness of the information is calculated . ranking, and validity calculating means for changing, the efficacy degree was calculated by the availability calculation unit, equal to or greater than a preset threshold value private data Detection means for detecting information stored in the base as information to be moved to a public database, and presentation means for presenting the information to be moved detected by the detection means together with the validity rank of the information It is characterized by this.

  Further, the information maintenance management apparatus of the information sharing system described above further comprises data movement means for moving and storing the information to be moved detected by the detection means to the public database. is there.

In order to achieve the above object, the method of information maintenance management of the information sharing system of the present invention is a step of connecting to a computer network and storing information to be disclosed in response to a user's inquiry in a public database. And a step of accumulating information not to be disclosed in a private database, and a step of searching the private database in parallel with a query of the user to the public database or a reference to the information referred to as the private database. If there is relevant information in the step of searching the private database in parallel as a query to the private database, the effectiveness frequency is calculated by adding the score corresponding to the validity ranking of the information, the in effectiveness counts and compares the pre Furiate obtained other information, the validity order of the information A step of further comprising the steps of detecting as information to be moved the information stored in the private databases efficacy frequency calculated is equal to or greater than a preset threshold value by the step in the public database to change the effective rank The step of presenting the information to be moved detected in the detecting step together with the order of validity of the information, and the step of moving and storing the information to be moved to the public database. It is a feature.

  According to the information sharing system of the present invention, when provisional information before publication is registered as non-public information, priority is given to information before publication with high demand from the contents of user operations on the published information. Since it is presented with a certain degree of time, disclosure of highly effective information is promoted.

(First embodiment)
Embodiments of the present invention will be described below with reference to the accompanying drawings. FIG. 1 shows a schematic block diagram of an embodiment of an information sharing system.

In FIG. 1, the information sharing system 10 includes a reliability prediction unit 1, an effectiveness calculation unit 2, a presentation unit 3, and a detection unit 4. In this embodiment, the information sharing system 10 includes an input / output unit 9 that receives input from a user and a storage unit 8 that stores information.

  The accumulating unit 8 accumulates information such as documents describing individual knowledge and know-how, computer software, and pictures. In addition, a history of user operations performed on each information is also accumulated.

  The input / output unit 9 includes a device that gives instructions to the system such as a keyboard and a mouse, and a device that presents information from the system such as a display to the user. In this embodiment, the storage unit 8 is a database on a computer, and the input / output unit 9 is a computer input device and a screen. However, in the present invention, the forms of the storage unit 8 and the input / output unit 9 are not limited. .

  The user uses the input unit 9 to give a request for information registration, search, presentation, etc. to the system. The system operates the storage unit 8 in accordance with the request, and presents the result to the user through the input / output unit 9.

  The reliability prediction unit 1 is a means for predicting the possibility of the accuracy of information stored in the storage unit 8 at a certain point in time from the interval at which the information was updated in the past. The validity calculation unit 2 is a unit that calculates the validity of the information at a certain point of time based on the user's operation on the information stored in the storage unit 8.

  The presenting unit 3 is a means for receiving the output results of the reliability predicting unit 1 and the effectiveness calculating unit 2 as input and determining whether or not the information needs to be reviewed and presenting the result to the user together with the reason for the determination. is there.

  FIG. 2 shows an example of the contents of information stored in the storage unit 8 in the information sharing system 10 according to the first embodiment. Each type of information includes an information identification number 21, an attribute 22, and an attribute value 23 as shown in FIG. The attributes include title, registered user ID, registration date, classification indicating information type (for example, know-how, document, software, image data, etc.), related information, search keyword, content, reference history, and the like. Each information is added with a history of operations such as reference, correction, and evaluation performed by the user.

  FIG. 3 is an example of a file for storing an evaluation list added to information. A list 32 of evaluations added to the information attribute 31 is accumulated as a table of dates, evaluators, and evaluation contents. FIG. 4 is an example of a file for storing a list of operations performed on information. A list 42 of operations performed by the user on the information attribute 41 is stored as date, user ID, and operation content information.

  In the example of FIG. 2, the information, the evaluation history for the information, and the operation history are stored on the same medium. However, in the present invention, the storage method, for example, information attribute information and information content (document, software, It is not limited whether the image data is stored in the same medium, and whether the information and the evaluation history for the information are stored in the same medium. For example, the content of information may be on another system connected via a network.

  5 and 6 show an embodiment of a user interface for a user to search and reference information via the input / output unit 9. In FIG. 5, the information search window 51 includes a search request input unit 52 and a main screen 53. The user inputs a search request for information into the system on the search request input unit 52.

  The search request is given a logical expression of keywords such as “Business trip report and (creation date = September 1998)”, and a natural language sentence such as “Show me an example of a business trip report created last September” It is done. The system searches for information satisfying the input search request from the information stored in the storage unit 8 and outputs the result on the main screen 53. FIG. 5 shows an output example of the search result. When there are a plurality of pieces of information that satisfy the search request, the system displays the titles of those pieces of information as a list. The target information is specified by the user selecting information to be referred to from the subject list.

  In FIG. 5, a subject list of a plurality of information is displayed on the main screen 53. FIG. 6 is a display example of a screen on which a user refers to the information of FIG. 5 through the information reference unit.

  The user can input an evaluation into the information presented by the system in response to the search request. FIG. 7 is an example of a user interface for a user to input an evaluation for information via the input / output unit 9 and the presentation unit 3. User. By instructing buttons 72 such as “very useful” and “useful” on the evaluation input window 71, an evaluation for information can be input. It is also possible to input a more detailed evaluation on the comment input form 73.

  Next, the presentation unit 3 of the information sharing system 10 in FIG. 1 detects information that needs to be reviewed from the values calculated using the results of the reliability prediction unit 1 and the effectiveness calculation unit 2 as scale factors, and the detected information is Explain the information maintenance and management methods to be presented together with the reason for review.

  FIG. 7 is a flowchart showing a flow of processing in which the presentation unit 3 presents information that needs to be reviewed to the administrator. First, the presentation unit 3 obtains the reliability x (i) of the information from the reliability prediction unit 1 for certain information (step S501).

  Hereinafter, a method in which the reliability prediction unit 1 predicts the reliability of information in the storage unit 8 will be described. Assume that the total number of information stored in the storage unit 8 is N, and the set of information is D = {d (1), d (2),..., D (N)}.

  Further, the number of times the review is performed for each element of D is C = {c (1), c (2),..., C (N)}.

  That is, the i-th element d (i) of D has been reviewed c (i) times by a certain time after registration.

Assume that the time (number of days) at which d (i) is updated is {t _i1 , t _i2 , t _i3 ,..., t _{ic (i)} , t _{ic (i) +1} }.

Here, _{tic (i) +1} represents the number of days that have elapsed since the last change. The default value of the review interval set by the administrator is TH.

The reliability predicting unit 1 obtains the number of times c (i) at which the data d (i) has been reviewed. When c (i) = 0, that is, when the review of the information d (i) has not been performed since registration, t _i1 / TH is set as the reliability x.

If c (i) ≠ 0, the next review interval is predicted from the previous review interval. For each j where 0 <j <c (i) +1, add p _i (j) × t _ij . This is TH (i). (1)
Here, p _i (j) is a weight parameter used for reliability prediction, and is represented by, for example, an equation of p _i (j) = a _i × j + b _i (2).

Further, the condition of p _i (1) + p _i (2) + p _i (3) +... + P _i (c (i)) = 1 is satisfied. In the present embodiment, the weight parameter is mapped on a single straight line equation defined by equation (2), but the method of determining the parameter is not limited.

For TH (i) obtained in (1), t _{ic (i) +1} / TH (i) is calculated, and this is defined as reliability x (i). The reliability x (i) has a value of 0 when the information is registered or reviewed, and then increases as time passes. Information whose reliability value exceeds 1 is information that has not been reviewed even after the review time expected by the system has passed.

  Next, the validity y (i) of the information is obtained from the validity calculation unit 2 (step S502). Hereinafter, a method by which the validity calculation unit calculates the validity of information will be described. The number of times the user has referred to the information d (i) stored in the storage unit 8 is r (i).

  In addition, the evaluation value e (i) of the information d (i) is calculated from the evaluation input by the user to the information in the screen example shown in FIG. In the example of FIG. 7, evaluations of “very useful”, “useful”, “insufficient information”, and “old information” can be input to the information. The value e (i) is obtained. Note that E = {e (1), e (2),..., E (N)}.

The operation performed by the user on the information d (i) is assumed to be {o _i1 , o _i2 , o _i3 ,..., O _{ir (i)} }. Here, r (i) is the number of operations performed on the information d (i). In addition, the time when each operation is performed is assumed to be {T _i1 , T _i2 , T _i3 ,..., T _{ir (i)} }.

The evaluation value v (o _ij ) is determined as follows for each operation o _ij .

v (o _ij ) = 1 (when o _ij is a reference)
v (o _ij ) = 2 (when o _ij is evaluated as “useful”)
v (o _ij ) = 3 (when o _ij is evaluated “very useful”)
When the current time is tt, the score e (i) of the information d (i) is calculated as follows.

e (i) = v (o _i1 ) / (tt−T _i1 ) + v (o _i2 ) / (tt−T _i2 ) +... + v (o _{ir (i)} ) / (tt−T _{ir (i)} ) (3)
Finally, the effectiveness y (i) is changed to y (i) = e (i) / AVE (E) (4)
It is calculated by the following formula.

  Here, AVE (E) is an average value of e (i) for all data stored in the storage unit 8. Effectiveness y (i) shows a high value when information is frequently referred to recently and high evaluation is obtained.

  A procedure for determining whether the detection unit 3 should review the information based on the reliability x and validity y of the information calculated by the method described so far will be described below.

The presentation unit 3 uses the reliability x (i) of the information and the validity y (i), and w (i) = (x (i) × x (i) + y (i) × y (i)) ^{1 / 2} ... The value of (5) is obtained (step S503).

  This w (i) is compared with a predetermined threshold value TH1 (S504). FIG. 9 shows a two-dimensional space with x (i) and y (i) as axes. Here, w (i) is a distance from the origin to a certain coordinate, and whether w (i) is larger than the threshold value TH1 results in whether or not it enters the inside of the circle 91 in the two-dimensional space. Information d (i) in which w (i) is plotted outside the circle is information that needs to be reviewed, and information d (i) plotted inside the circle is information that does not need to be reviewed. It is done.

  If w (i) is smaller than the threshold TH1, the system checks whether the reliability x is larger than 1 (S505).

  If it is greater than 1, the information d (i) has exceeded the review period predicted by the system. Moreover, since w (i) is small, it is considered that the effectiveness of d (i) is not so high. Therefore, the system performs processing such as lowering the priority of retrieval of information d (i) or automatically deleting it (S509).

  When w (i) is larger than the threshold value TH1, the presentation unit 3 obtains a reason for reviewing the information d (i) (S506).

  First, it is checked whether or not the reliability x (i) is larger than 1. When x (i)> 1, the information d (i) is information that has exceeded the review period predicted by the system and has been evaluated to some extent by the user. Therefore, the presentation unit 3 searches the past evaluation history added to the information d (i), and obtains the usage status such as “the number of reference is large” or “having a high evaluation”. This determines the reason that “the review period has passed but the evaluation is high”.

  If x (i) <1, but w (i) is greater than TH1, effectiveness y (i) indicates a high value, that is, information d (i) is recently high. Since it can be said that the evaluation has been obtained, the presenting unit 3 determines the reason that “the review deadline is not reached, but the content needs to be confirmed because the evaluation is high”. It is also possible to add the result of the reference count and the evaluation value obtained from the evaluation history.

  Next, the presentation unit 3 requests the administrator to perform an operation for correcting or deleting the content (S507). Here, the reason determined in the previous step S506 is also presented.

  FIG. 10 is a display example of a screen requesting the administrator to correct / delete information. On the information management request window 100, an attribute 101 such as ID, title, author, etc. of information to be corrected / deleted is displayed, and a reason 102 for requesting deletion / correction of information is displayed below it. Yes. When the administrator designates the information deletion button and the correction button 103, the information can be corrected and deleted.

Further, by instructing the notification cycle change button 104, it is possible to give an instruction to extend the notification cycle by the presentation unit 3, that is, to reduce the value of a _i in equation (2).

  FIG. 11 shows another display example of a screen requesting the administrator to correct / delete information. In this example, w (i) for all information is calculated and presented in descending order. In addition to the title of each information, the value of w (i) for the information and the reason why correction / deletion is necessary are presented. As a result, a list of information that needs to be corrected or deleted as soon as possible can be presented to the administrator in the order in which the review is necessary.

  In the example of FIG. 11, w (i) is displayed as a numerical value. However, in the present embodiment, the display method is not limited to this method. For example, it may be displayed as a bar graph instead of a numerical value, or an arrow may be displayed as a vector on the x (i), y (i) plane.

  Further, the reason for determining that correction is necessary is not limited to the method of FIG. For example, instead of displaying the reason directly, a method of changing the character of the title part or the above bar graph-like display depending on the type of reason may be considered.

  Next, the presentation unit 3 confirms that the reliability prediction unit 1 and the effectiveness calculation unit 2 have performed operations related to the information after correcting or deleting the content of the information according to the operation from the administrator. Notice.

The reliability prediction unit 1 changes parameters used for reliability prediction in accordance with the notification content (S508). In the present embodiment, the slope a _i and the initial value b _i are changed in the weight parameter determination formula shown in equation (2), p _i (j) = a _i × j + b _i . If there is an instruction to shorten the notification cycle, the value of a _i is increased, and if it is instructed to increase the value, the value of a _i is decreased. Accordingly, the value of b _i is adjusted so that the sum of p _i (j) becomes 1.

  Although the method for obtaining the reliability x (i) and the validity y (i) of the information has been described in the explanation of steps S501 and S502 in the flowchart of FIG. 8 of the present embodiment, the reliability and validity are improved in the present invention. The method to obtain is not limited to this.

For example, in the process of determining x (i) in step S501, the formula (2) for determining the weight parameter is set to p _i (j) = 1 / c (i), so that the average of the review intervals so far is It can be a predicted value of the review interval.

  In addition, when calculating TH (i) in (1), TH (i) is not added for all j satisfying 0 <j <c (i) +1, but in total for a predetermined constant m times. It is also possible to ask for

Also, review the number of times X = {1,2, ..., c (i)} of the information d (i) interval between each revision operation as _{Y = {t i1, t i2} , ..., t ic (i)} relation Can be approximated as a regression line Y = AX + B, the time interval until c (i) + 1st review is predicted using the parameters A and B of this line, and the value is used as TH (i). It is done.

Even in the process of obtaining y (i) in step S502, the score for each evaluation content when determining the value of v (o _ij ) is not limited to the numerical value of the present embodiment. Further, equation (3) for obtaining e (i) is not limited to this calculation method. Further, in the equation (4), normalization is performed with the average value of e (i) for all data, but it is also conceivable that this is also normalized with a predetermined constant instead of the average value.

Also, the method of inputting an evaluation for information by the user shown in FIG. 7 is not limited to the method of this embodiment with respect to the type of evaluation to be input or the input method. Further, the evaluation input interface as shown in FIG. 7 is not essential for realizing the present claims. For example, it is possible to calculate the effectiveness from only the number of information references without using the interface for evaluation input. In that case, V (o _ij ) used in the expression (3) may be set to a constant.

  This content is not limited even in the part shown in step S508 in which parameters for calculating reliability and effectiveness are changed based on the operation result. Whether or not to execute step S508 is not limited in the present invention.

(Second embodiment)
FIG. 12 is a schematic block diagram of an information sharing system that implements an information maintenance / management system according to another embodiment of the present invention. 12, the information sharing system includes an information registration unit 121, an information reference unit 122, a document analysis unit 123, a reliability determination unit 124, a review information detection unit 125, a detection information presentation unit 126, a public information storage unit 127, and a document analysis. It comprises knowledge 128, vocabulary knowledge 129, and input / output unit 1210. The information registration unit 121 generates information such as documents, computer software, and pictures describing knowledge and know-how and stores the information in the public information storage unit 127.

  The information reference unit 122 is for the user to search, reference, and use information stored in the public information storage unit 127. The document analysis unit 123 analyzes the information content stored in the public information storage unit 127 and understands the description content. The document analysis knowledge 128 describes knowledge necessary for document analysis, such as syntax analysis knowledge, and is referred to during document analysis. The vocabulary knowledge 129 describes knowledge about natural language sentence expressions and things, and the reliability determination unit 124 is based on the document analysis result and vocabulary knowledge, and the reliability of the document that the registration information is valid or useful. Determine gender.

  The review information detection unit 125 checks whether or not the information detected by the reliability determination unit needs to be updated, and detects information that needs to be reviewed. When presenting the review information, the detection information presentation unit 126 presents information using the document analysis result as the reason why the review is necessary.

  13 and 14 show information contents stored in the public information storage unit in such an information sharing system. As shown in FIG. 13, each type of information includes an information identification number 130, an attribute 131, and an attribute value 132. Attributes include title, publisher, publication date, classification representing know-how, software and painting, business classification, classification representing topic and other information categories, related information, keywords for search, contents, reference history, etc. is there.

  FIG. 13 is an example of know-how information. The title is a lecture holding procedure, the publisher is Taro Yamada, the publication date is October 1, 1995, the type is know-how, the classification is visitor service, and the related information is a request document. (Information identification number # 200), list of academics (# 350), keywords indicate lectures, visits, university professors and agenda, and the contents of the procedures (1) to (7) is described. Furthermore, as an additional history of additional information, an identification number stored in a semantic relationship storage unit that stores the semantic relationship between additional information and public information is described. In this example, / log / # 24, / log / # 38, / log / # 105.

  FIG. 14 shows an example of computer software. The title is a budget management program, the publisher is Momoko Yoneyama, the release date is May 10, 1993, the type is software, the classification is budget management, and related information is arranged. List of requests, keywords are budget, purchase schedule, overtime, contents are program files / etc / yosan-kanri, additional history is / log / # 60, / log / # 62, / log / # 120. .

  Next, a method for inputting information via the information registration unit 1 will be described. FIG. 15 is a diagram illustrating an example of a user interface of the information registration unit. The information registration window 151 includes a main screen 152 and a control panel 153.

  The [Classification designation] button 154 on the control panel 153 is for designating the classification of information. The classification, which is one of the attributes of information stored in the information storage section, also indicates the business classification or information category to which the information belongs. A user who accumulates information designates a classification related to the information via the information registration unit, so that the information is accumulated in the information accumulation unit in association with the classification. When an information publisher (hereinafter referred to as a user) selects the [Specify Classification] button 154 on the control panel 153, a conceptual diagram of a classification system related to information is displayed on the main screen 152 in a network form, for example. The The user associates the classification with the information to be registered by selecting one or more classifications from the classifications displayed on the window 152.

  [Information input] button 155 on control panel 153 is used to instruct to switch to a mode for inputting public information. When the user selects the [information input] button 155 on the control panel 153, the main screen 152 is switched to a screen for inputting information contents.

  FIG. 16 shows an embodiment when the user inputs text format information on the information registration window. The user inputs the title, type, format, and content of the information to be entered and published. The contents can be input directly, or by specifying a specific file name that has already been edited, the contents of the file can be disclosed as information. In this example, the format of information disclosed by the user is text, but the information registered in the information storage unit in the information sharing support system is not limited to text format, and still images, moving images, audio, programs It is also possible to share information such as the execution format by specifying the file name.

  Next, a method for searching and referring to public information stored in the information storage unit via the information reference unit 122 will be described.

  17 and 18 show an embodiment of a user interface for searching and referring to information through the information reference unit. In FIG. 17, the information search window 171 includes a main screen 172 and a control panel 173. An information referrer (hereinafter referred to as a user) designates search target information by selecting a [classification designation] or [keyword designation] button on the control panel.

  For example, by instructing the [Specify Classification] button 174, a conceptual diagram representing a classification system of target information is displayed on the main screen 172, and the corresponding classification is selected. When the user designates the [Execute Search] button 175 on the control panel 173 after designating a category or keyword, information related to the designated category or set of keywords is retrieved. When there are a plurality of search results, the information reference unit displays a subject list of each information, and the user selects information to be referred to from the subject list, thereby specifying target information. In FIG. 17, a plurality of information subject lists are displayed on the main screen 172. FIG. 18 is a display example of a screen on which the user refers to the information of FIG. 17 via the information reference unit.

  Next, in such an information sharing system, in order to update the document content when the registered document content is out of date or incorrect, the document content is analyzed, a document that needs to be reviewed is detected, The flow of processing presented to the system administrator will be described.

  FIG. 19 shows the contents of the document analysis knowledge 128 that is referred to when the document analysis unit 123 analyzes the document. FIG. 19 shows document analysis knowledge used to detect, for example, obsolete document content. 190 is knowledge for classifying the description of the document content. The knowledge is composed of an expression classification 192 and an expression set 193. The expression classification: : (Sub-classification): defined as an expression set.

  For example, among the classifications representing “time”, as expressions representing “period”, “from <date> (expression pattern representing time expression) to <date>”, “between <date>”, “<any>” (Unspecified expression pattern) Expressions such as “period” and “when” are defined. Reference numeral 191 denotes a definition of an expression pattern described in the knowledge described in the knowledge 190, and is composed of an expression pattern identifier 194 and definition contents 195.

  For example, the expression <date> matches any or all of the three expressions "<num> year", "<num> month", and "<num> day" (<num> is a number). (() Indicates an expression group, [] + indicates that it matches one or more expressions in the expression group), <num> months ”, or <num> weeks” Defines matching.

  The document analysis unit 123 analyzes the document with reference to such document analysis knowledge.

First, morphological analysis is performed on the registered document, and after the document is decomposed into morphemes, it is checked whether each morpheme matches the expression pattern of the document analysis knowledge 128. If there is a match, the expression classification is added as additional information at the position where the matched expression appears in the document.

  FIG. 20 shows, for example, the contents of vocabulary knowledge 129 that is referred to when recommending information is presented at a certain time or when frequently searched information is estimated. Vocabulary knowledge is described as a set of terms 201 and constraint descriptions that appear in the document. In this case, for example, the description about Hatsumode indicates that it is strongly related to December. For example, since it is often referred to in December, the content must be reviewed before December. Recommended information in December It can be used as a presentation. In addition, the restrictions described in the vocabulary knowledge include urgency such as whether the term is always related to business or is a term related to frequently referred documents.

  The reliability determination unit 124 determines the reliability of registered contents with reference to the document analysis result and vocabulary knowledge. Information with low reliability is, for example, old information. Therefore, determining whether information is stale is one example of determining reliability. The method for determining whether the information is old includes, for example, an expression whose expression classification is time, money, person / organization, document, address (address, telephone number, URL, etc.) as an analysis result of the document analysis unit 123. Document. In response to this result, the review information detection unit 125 presents a document that needs to be reviewed to the system administrator and prompts correction.

  FIG. 21 shows a processing flow of the reliability determination unit 124, the review information detection unit 125, and the detection information provision unit 126. First, the document is analyzed (S211), and the expression classification is extracted (S212). In the case of this embodiment, since it is an example of determining whether the content is old or not, it is checked whether or not the expression classification is likely to change over time (S213). If there is a corresponding expression classification, it is stored in the corresponding document list (S214). If not, the next document is inspected (S215).

  Next, vocabulary knowledge is referred to for a document having the corresponding expression classification, and it is checked whether there is a description that matches the vocabulary in the document (S216). In the case of this embodiment, contents highly relevant to the current time are inspected (S217). For example, if the current time is December, it is checked whether or not it includes vocabulary such as “year-end adjustment”, “year-end party”, “banquet”, etc., in which “period = December” is set.

  If there is a corresponding document, the priority of the presentation order is increased (S218). This process is performed for all documents, and finally the documents accumulated in the corresponding document list are rearranged in order of priority and presented (S219).

  FIG. 22 is an example of the display screen of the detection information providing unit 126. A list of documents detected by the review detection unit 125 and required to be reviewed is displayed in order of priority.

  Reference numeral 220 denotes a document title, 221 denotes an expression classification indicating the reason for correction, 222 denotes a priority, 223 denotes an excerpt of a review part (a part where an expression pattern of the expression classification appears). For example, in the document “Short-range daily allowance”, the expression classification is a monetary amount and the priority is “always urgent”. Always urgent indicates that the urgency is high because it is frequently referenced regardless of the season.

  On the other hand, the “communication ticket purchase method” is always referred to, but it is “always” because the frequency is low. Also, in “Short-range daily allowance”, since the portion where “daily allowance is 1000 yen” is an expression pattern corresponding to the expression classification, the corresponding portion is displayed in reverse.

  FIG. 23 and FIG. 24 are screen examples in which the review document is presented to the system administrator. The portion that needs to be reviewed, that is, the portion whose expression classification matches the item that needs to be corrected is highlighted (230). In this case, it is urged to review the portion where the money expression of 1000 yen appears. On the other hand, the system administrator can correct it by himself, but in this example, “correction request” is selected from the menu 231 to prompt the information publisher to make corrections. Then, as shown in FIG. 24, an additional information input window 242 is displayed, in which the contents of the correction request are described and transmitted to the information publisher.

Embodiment 1

  Next, an embodiment of the present invention will be described. FIG. 25 shows a functional configuration of the information maintenance / management method according to the present embodiment.

  This information maintenance and management method is used at customer consultation desks that answer customer inquiries. From the history of individual inquiries, a generalized assumption inquiry case is created for the contents of frequent inquiries. Is published as public document data that can be directly searched.

  The public database 251 stores information that can be disclosed and directly searched by customers. Similar to the first embodiment, a keyword logical expression and a natural language sentence can be searched as a search request. For example, in response to a search request for a natural language sentence, an angle between an independent word vector extracted by morphological analysis and an independent word vector of each information in the public database 251 is obtained by using an information search method based on a vector space method. Enable to search similar information as similarity. For example, it is determined that the closer the angle is to 0, the more similar. Further, if a keyword logical expression is set, a document satisfying the logical expression is searched.

  The non-public database 252 stores inquiries sent to customer consultation counters from routes other than the public database 251 and the contents of responses from the customer counseling staff in response to the inquiries.

  The validity calculation unit 253 applies the content of access to the public database 251 by the user to the private database 252 and obtains the validity of each information in the private database 252 based on the result.

  The detection unit 254 detects information to be moved to the public database 251 in the private database 252 based on the validity obtained by the validity calculation unit 253. The presentation unit 255 presents the information detected by the detection unit 254 together with the result of the effectiveness calculation unit 253.

  Hereinafter, this embodiment will be specifically described. FIG. 26 shows an example of information stored in the public database 251. Here, assumed inquiry cases that frequently occur with respect to a certain product group are accumulated. FIG. 27 shows an example of information stored in the private database 252. Here, the contents of the inquiry sent to the customer consultation counter by telephone or FAX and the contents of the response of the person in charge of the inquiry are stored.

  The information in the private database 252 is log information of inquiries sent to the customer consultation desk, and is not necessarily high-quality information that is meaningful to other customers. Since disclosure of low-quality information leads to degradation of service for the customer consultation service, the non-public database 252 is not disclosed to the customer here.

  Now, it is assumed that an inquiry shown by reference numeral 281 in FIG. 28 is input from the user to a search unit different from the present invention. The search unit 254 searches the public database 251 by a known search method and presents the result to the user. Here, the query that is searched for the public database 251 in this way is input to the effectiveness calculation unit 253.

  The validity calculation unit 253 having input FIG. 28 calculates the validity of the information in the private database 252 according to FIG. Here, the search performed by the effectiveness calculation unit 253 is assumed to be a conventionally known method. The input to the effectiveness calculation unit 253 may be other than an inquiry to the public database 251.

  For example, when the user browses the information in the public database 251, the content of the browsed information such as 282 may be used as the input to the validity calculation unit 253. In this case, the effectiveness calculation unit 253 may perform a search called a similar document search. That is, instead of a user inputting a search request for a natural language sentence, the browsed document may be searched as a search request.

  For example, SCORE in FIG. 29 is obtained as follows. SCORE = (11−RANK) * s (11−RANK). Here, RANK is a rank in the non-public database in the search by the effectiveness calculation unit 253. Further, s (n) is a step function, and takes a value of s (n) = 1 when n> 0 and s (n) = 0 when n ≦ 0.

  That is, the information that was ranked first as a result of the search has a SCORE of 10, and becomes 9, 8,. The SCORE of the 10th information is 1, and the SCORE is 0 for all the information of the 11th and lower.

  For example, if 271 in FIG. 27 is second in the search by the effectiveness calculation unit 253, the SCORE is 9, and the effectiveness of 271 is 32 according to FIG. As a result, the information in the non-public database 252 becomes as 301 in FIG.

  SCORE may be determined by other criteria. For example, it is conceivable to use a method for obtaining not only the rank but also the score and the similarity between the search request and the information as a reference for the SCORE.

  Receiving the result of the effectiveness calculation unit 253, the detection unit 254 detects information to be moved to the public database 251 in the private database 252 according to FIG.

For example, THRESHOLD in FIG. 31 has a value such as THRESHOLD = 30.

  In other words, the detection unit 254 detects information in the private database 252 corresponding to the SCORE being selected as the first place three times for the input query as information to be moved to the public database 251.

  For example, 301 in FIG. 30 has an effectiveness of 32 and is larger than THRESHOLD = 30 here, and is detected as information to be moved to the public database 251 in FIG. THRESHOLD may be determined based on other criteria. For example, another constant may be used, and it is also conceivable to determine THRESHOLD by the average of SCORE at the number of times of input.

  The presenting unit 255 presents the information detected by the detecting unit 254 to the person in charge of the customer consultation counter together with the validity calculated by the validity calculating unit 253. FIG. 32 shows an example of presentation by the presentation unit 255.

  Thereby, the person in charge of the customer consultation counter can use the contents presented by the presentation unit 255 as a clue when transferring information in the non-public database 252 to the public database 251.

  In this embodiment, the public database 251 and the private database 252 may be physically the same database. The public database 251 and the private database 252 can be logically distinguished by holding a flag indicating whether or not each data is disclosed.

  In such a configuration, whether the search for the public database 251 and the search performed by the validity calculation unit 253 for the non-public database 252 are the same, and the search result is presented to the user depending on whether or not to disclose. Should be distinguished. Since searches for two types of data can be completed at once, the burden on the system can be reduced by reducing the number of times of search processing. At this time, RANK when obtaining the above SCORE may be the order of the information of the private database in the information of both the public database and the private database.

  Further, when the public database 251 and the private database 252 are physically the same database, when the administrator moves the information of the private database 252 to the public database 251 in accordance with the presentation contents of the presentation unit 255. As a process, it is only necessary to change the flag of whether to disclose or not, so the load on the system is also reduced.

  Alternatively, inquiries from users and information that has been viewed, which are inputs to the effectiveness calculation unit 253, can be stored as a log, and maintenance and management can be performed in batches using this information. By doing so, for example, processing can be performed collectively at night when there are few inquiries from the user, so that the load on the system related to maintenance and management is reduced when there are many inquiries from the user.

  Further, the processing in the embodiment of the present invention can be realized by a computer-executable program, and this program can be realized as a computer-readable storage medium.

  As the storage medium in the present invention, a program such as a magnetic disk, floppy (registered trademark) disk, hard disk, optical disk (CD-ROM, CD-R, DVD, etc.), magneto-optical disk (MO, etc.), semiconductor memory, etc. As long as the storage medium can be stored and read by the computer, the storage format may be any form.

  Further, an OS (operation system) operating on the computer based on an instruction of a program installed in the computer from the storage medium, database management software, MW (middleware) such as a network, and the like for realizing the present embodiment A part of each process may be executed.

  Furthermore, the storage medium in the present invention is not limited to a medium independent of a computer, but also includes a storage medium in which a program transmitted via a LAN, the Internet, or the like is downloaded and stored or temporarily stored.

  Further, the number of storage media is not limited to one, and the processing in the present embodiment is executed from a plurality of media, and the configuration of the media may be any configuration included in the storage media in the present invention.

  The computer according to the present invention executes each process according to the present embodiment based on a program stored in a storage medium, and includes a single device such as a personal computer or a system in which a plurality of devices are connected to a network. Any configuration may be used.

  The computer in the present invention is not limited to a personal computer, but includes a processing unit, a microcomputer, and the like included in an information processing device, and is a generic term for devices and devices capable of realizing the functions of the present invention by a program. .

The block diagram which shows the structure of a 1st Example. An example of information stored in the storage unit. An example of a screen where a user searches for information. An example of a screen where a user searches for information. An example of a screen where a user searches for information. An example of a screen where a user searches for information. A screen example in which a user adds evaluation information to information. The figure which shows the process sequence of a determination part. The figure used for demonstrating the determination method of a determination procedure. Example screen requesting the administrator to review information. Another example of a screen requesting the administrator to review information. Functional configuration of the second embodiment. An example of accumulated information. An example of accumulated information. Example registration interface. An example of a registration operation. Reference interface example. Reference interface example. An example of document analysis knowledge. An example of vocabulary knowledge. Flow of processing of the reliability determination unit. An example of a detection information display screen. Example administrator interface. Example administrator interface. The function block diagram of embodiment of this invention. An example of information in a public database. An example of information in a private database. An example of input to the effectiveness calculation unit. The example of the flow of a process of an effectiveness calculation part. An example of information with modified effectiveness. The example of the flow of a process of a detection part. The example of the presentation content by a presentation part.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Reliability prediction part 2 ... Effectiveness calculation part 3 ... Presentation part 4 ... Detection part 8 ... Accumulation part 9 ... Input / output part 10 ... Information sharing system

Claims (3)

A public database that is connected to a computer network and that publishes the accumulated information in response to a user's inquiry, and a private database that accumulates information that is not subject to disclosure;
The user's query to the public database or the referenced information is searched in parallel as a query to the private database, and if there is applicable information, the validity ranking of the information is determined. Effectiveness calculation means for calculating the effectiveness frequency obtained by adding the corresponding score, comparing this with the effectiveness frequency of other information allocated in advance, and changing the effectiveness ranking of the information,
Detection means for detecting information stored in a private database whose effectiveness frequency calculated by the validity calculation means is equal to or greater than a preset threshold as information to be moved to the public database;
Presenting means for presenting the information to be moved detected by the detecting means together with the validity rank of the information;
An information maintenance management apparatus for an information sharing system, comprising: 2. The information maintenance management apparatus for an information sharing system according to claim 1, further comprising data movement means for moving and storing the information to be moved detected by the detection means to the public database. Connecting to a computer network and storing in a public database information for publication in response to user inquiries;
Storing non-public information in a private database;
A user query to the public database, or information referred to, searching the private database in parallel as a query to the private database;
If there is relevant information in the step of searching the private database in parallel as a query to the private database, the effectiveness frequency obtained by adding the score corresponding to the validity ranking of the information is calculated in advance. Changing the effectiveness ranking of the information compared to the effectiveness frequency of the other information assigned to it,
Detecting the information accumulated in the private database whose effectiveness frequency calculated by the step of changing the validity ranking is equal to or greater than a preset threshold as information to be moved to the public database;
Presenting the information to be moved detected by the detecting step together with the validity rank of the information;
Moving and storing the information to be moved to the public database;
A method for information maintenance management of an information sharing system, comprising:

Images