JP2009187077A - Recommendation information providing apparatus, recommendation information providing method, recommendation information providing system, and recommendation information providing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To recommend an appropriate person according to a similar situation in which a user has been involved in the past. <P>SOLUTION: A recommendation information providing apparatus has: a history information storage means 1 for storing feature data of application data in association with data identification information; a relation information storage means 2 for storing information related to the application data as relation destination information with data identification information as relation source information and storing a relation class indicating relation between a relation source and a relation destination; a similarity information decision means 3 for calculating similarity between new application data and the feature data to output data identification information corresponding to feature data of high similarity together with the similarity; and a recommendation information decision means 4 for searching the relation information storage means 1 based on the output data identification information, taking a value with similarity varied as a recommendation degree of the relation destination information based on relation between a relation class corresponding to relation destination information matching a recommendation request item and the user of an application of the relation class, and taking the relation destination information of high recommendation degree as recommendation information. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a recommendation information providing apparatus, a recommendation information providing method, and recommendation information provision for providing recommended information to a user in order to support the user's information input to an application for communication in a computer network environment represented by the Internet. The present invention relates to a system and a recommendation information providing program.

  When communicating with others via e-mail, telephone, video conferencing, etc. via a computer network, it is necessary to enter the e-mail address or telephone number of the other party. When performing these inputs, the user usually obtains the other party's information by referring to an address book associated with the application realized by the application program or performing a search from a person name, and inputs the information.

  An example of a technique for supporting information input activities and reducing the burden on the user is described in Patent Document 1. In the invention described in Patent Document 1, each destination when an email is transmitted at the same time in the past is held as a group, and when the user inputs one destination, the email is transmitted simultaneously with the input destination. The other destinations are searched from the past history and recommended as input candidates.

  Further, in the invention described in Patent Document 2, the destination name and subject of an email sent in the past are stored in the storage means in association with each other, and similar to the information newly entered in the subject column of the email. Information is retrieved from the storage means, and a destination name corresponding to a subject including information similar to the information input in the subject column is extracted and recommended, thereby reducing the burden required for the user's destination setting.

  Further, in Patent Document 3, a destination of an electronic mail transmitted in the past and a set of vocabulary considered to be important in the main body of the electronic mail are stored in a storage unit in association with each other. When created, the similarity between one or more vocabularies in the body of a new e-mail and the set of vocabularies stored in the storage means is calculated, and stored when the similarity is higher than a predetermined value. A method for recommending a destination corresponding to a set of vocabulary in the means is described.

JP 2004-310599 A (paragraphs 0015-0018) JP 11-272579 A (paragraphs 0012-0014) Japanese Patent Laid-Open No. 11-46212 (paragraphs 0045-0051)

  However, in the invention described in Patent Document 1, a destination to which frequent e-mails are transmitted is preferentially recommended, or a first input destination is a destination that is frequently used as a destination. In this case, there is a problem that a large number of destinations are recommended. In general, e-mails or the like have different destinations depending on the situation, but the invention described in Patent Document 1 cannot make a recommendation according to such a situation. For example, although it is sent only once a year, it is not possible to accurately recommend a person to be added to the destination in certain situations.

  Moreover, in the invention described in Patent Document 2, there is a problem that the destination name is not recommended when the subject field of the e-mail is not similar. In general, the subject of an email is at most 10-20 characters. In order to make it similar within that range, it is necessary to store the subject of past e-mails, which is a heavy burden on the user.

  In the inventions described in Patent Documents 1 and 2, the destination is recommended using the user's own past transmission history, but there is a problem that the recommendation including the transmission history of others is not made. Even if the user is inexperienced, it is often the case that others are already experienced. In some cases, it is possible to make a more appropriate recommendation in consideration of other people's past situations. However, since the recommendation using the information of other people is not made, appropriate information is not recommended when the user is inexperienced.

  In order to solve the above-described problem, an appropriate person is not recommended by simply changing the methods described in Patent Documents 1 and 2 so as to integrate a database for storing other person's information. For example, the boss for a certain person A may be an employee of another department for another person B, and the destination of another person B is an inappropriate destination as a destination for the certain person A. There are many. If the methods described in Patent Documents 1 and 2 are changed so that a database for storing other people's information is also integrated, a destination that is inappropriate for a person A is recommended as a destination.

  In the method described in Patent Document 3, when an e-mail text including only a vocabulary that is not similar to a set of vocabularies associated with the destination of the person B is created in the storage unit, The person B's destination is not recommended when he / she wants to send it.

  As described above, in the methods described in the above-mentioned documents, appropriate information according to similar situations that the user has performed in the past is not recommended. The reason is that there is no mechanism for judging a similar situation in the past and making a recommendation based on the situation.

  In addition, when the user has never experienced a similar situation in the past, appropriate information is not recommended even if another person has experienced a similar situation. The reason is that it does not have a mechanism to handle history information of multiple persons in an integrated manner, search both the past history of one's past and the past history of another person at the same time, and use the result for determining recommendation information. It is.

  In addition, when recommending appropriate information to the user based on the experience of others, if the recommendation is based on the history performed in another department or another organization, the appropriate information is not recommended. The reason is that it does not have a mechanism for converting transmission information of other departments or organizations into persons of their own departments or organizations.

  Therefore, the present invention provides a recommendation information providing apparatus, a recommendation information providing method, a recommendation information providing system, and recommendation information that can recommend an appropriate person according to a similar situation that the user has performed in the past in a system that recommends information such as destination information. The purpose is to provide the provided program.

  Another object of the present invention is to provide a recommended information providing device and a recommended information providing device that can recommend an appropriate person based on the past experience of another person even if the user has never experienced a similar situation in the past. It is to provide a method, a recommended information providing system, and a recommended information providing program.

  Still another object of the present invention is to recommend an appropriate person to a user based on the experience of another person, based on the history of other departments or other organizations, To provide a recommendation information providing apparatus, a recommendation information providing method, a recommendation information providing system, and a recommendation information providing program capable of recommending an appropriate person of an organization.

  The recommendation information providing apparatus according to the present invention uses, as history information, feature data that characterizes the application data extracted from application data transmitted from a processing apparatus that executes processing according to application software, to identify data attached to the application data. As history information storage means for storing information corresponding to the information, as the relation information, the data identification information of the application data transmitted from the processing device is used as the relation source information, and the information related to the application data is stored as the relation destination information. Similarity between the relationship information storage means for storing the relationship type indicating the relationship between the relationship source and the relationship destination, the application data transmitted together with the recommendation request item from the processing device, and the feature data stored in the history information storage means Calculate the relative A similarity information determination unit that outputs data identification information corresponding to feature data having a high degree of similarity together with the degree of similarity, and searches the stored contents of the relation information storage unit based on the data identification information output by the similarity information determination unit, and requests a recommendation Based on the relationship type corresponding to the related party information that matches the item and the relationship with the user of the application of the related type, the value that changes the similarity is used as the recommended level of the related information, and the recommendation level is relative And recommendation information determining means using recommendation information as high-level related party information.

  In the recommended information providing method according to the present invention, the feature data that characterizes the application data extracted from the application data transmitted from the processing device that executes the process according to the application software is used as the history information. The data stored in the history information storage means in association with the information, and as the relation information, the data identification information of the application data transmitted from the processing device is used as the relation source information, and the information related to the application data is stored as the relation destination information. In addition, processing for storing the relationship type indicating the relationship between the relationship source and the relationship destination in the relationship information storage means, application data transmitted together with the recommendation request item from the processing device, and feature data stored in the history information storage means Similarity with Processing for calculating and outputting data identification information corresponding to feature data having a relatively high degree of similarity together with the degree of similarity, searching the stored contents of the relational information storage means based on the output data identification information, and a recommendation request item Based on the relation type corresponding to the relation destination information that matches the relation destination information and the relationship with the user of the application of the relation type, the value that changes the similarity is set as the recommendation degree of the relation destination information, and the recommendation degree is relative And a process of using recommended information as related information.

  In the recommendation information providing system according to the present invention, the server associates the feature data characterizing the application data extracted from the application data transmitted from the processing device as the history information with the data identification information attached to the application data. The history information storage means for storing and the relationship information, the data identification information of the application data transmitted from the processing device as the relationship source information, the information related to the application data as the relationship destination information, and the relationship source and the relationship A relationship information storage means for storing a relationship type indicating a relationship with the destination; a degree of similarity between the application data transmitted together with the recommendation request item from the processing device and the feature data stored in the history information storage means; Corresponds to feature data with relatively high similarity Similar information determination means for outputting the data identification information together with the similarity, and the stored contents of the related information storage means based on the data identification information output by the similarity information determination means, and the related destination information that matches the recommendation request item Based on the corresponding relationship type and the relationship with the user of the application of the relationship type, the value that changes the similarity is set as the recommendation level of the related information, and the related information with a relatively high recommendation level is recommended. And recommending information determining means as information.

  The recommended information providing program according to the present invention is characterized in that feature data characterizing the application data extracted from application data transmitted from a processing device that executes processing according to application software is attached to the application data as history information. Processing related to the data identification information stored in the history information storage means, and as related information, the data identification information of the application data transmitted from the processing device is used as relation source information, and information related to the application data is related destination information Stored in the relationship information storage means, the application data transmitted together with the recommendation request item from the processing device, and the history information storage means. Processing for calculating similarity with feature data, outputting data identification information corresponding to feature data having relatively high similarity together with similarity, and stored contents of relation information storage means based on output data identification information Based on the relationship type corresponding to the related party information that matches the recommendation request item and the relationship with the user of the application of the related type, the value obtained by changing the similarity is used as the recommended degree of related information. , And a process of using related information having a relatively high recommendation level as recommendation information.

  According to the present invention, even when the user has never experienced a similar situation in the past, there is an effect that an appropriate person can be recommended based on the past experience of other persons.

  FIG. 1 is a block diagram showing an outline of the present invention. As shown in FIG. 1, the recommendation information providing apparatus according to the present invention is characterized by characterizing application data extracted from application data transmitted from a processing apparatus 5 that executes processing according to an application (application software) 150 as history information. History information storage means 1 for storing data in association with the data identification information attached to the application data; and as related information, the data identification information of the application data transmitted from the processing device 5 is used as relation source information. Information related to data is stored as related party information, relation information storage means 2 for storing a relation type indicating a relation between the relation source and the relation destination, and application data transmitted together with the recommendation request item from the processing device 5 , History information storage hand The similarity information determining means 3 for calculating the similarity to the feature data stored in 2 and outputting the data identification information corresponding to the feature data having a relatively high similarity together with the similarity, and the similarity information determining means output The stored contents of the relationship information storage means 1 are searched based on the data identification information, and based on the relationship type corresponding to the relationship destination information that matches the recommendation request item and the relationship with the application user of the relationship type, A recommendation information determination unit 4 is provided which uses a value obtained by changing the similarity as a recommendation level of related party information and uses related party information having a relatively high recommendation level as recommendation information.

Embodiment 1. FIG.
FIG. 2 is a block diagram showing a first embodiment of the recommendation information providing apparatus according to the present invention. In the first embodiment, the server 100 corresponding to the recommended information providing apparatus includes a storage unit that stores information and a calculation unit that performs calculation processing. The server 100 recommends information to a processing device such as a terminal device such as a personal computer equipped with the application 150. Hereinafter, it is expressed that the processing apparatus equipped with the application 150 transmits data or receives data when the application 150 transmits data or receives data.

  Storage means in the server 100 includes a history information storage unit 101, a person information storage unit 102, a relationship information storage unit 103, and a conversion rule storage unit 104. The computing means in the server 100 includes a history / relationship information conversion unit 111, a feature word extraction unit 112, a similar information estimation unit 113, and a recommendation information determination unit 114.

  That is, the relationship between history information storage unit 101 that stores history information such as e-mails and files together with feature words, person information storage unit 102 that stores person information, and information such as e-mails and senders and attached files. And a storage device having a conversion rule storage unit 104 that holds a rule for converting data transmitted from the application 150 into history information and relationship information, and receives data from the application 150, A history / relationship information conversion unit 111 that converts history information and relationship information into history information and relationship information based on the conversion rule acquired from the conversion rule storage unit 104 and outputs the history information and relationship information to the relationship information storage unit 103; A feature word extraction unit 112 that analyzes the contents of the mail and extracts a feature word, and an application 50, a similar information estimation unit 113 that estimates past similar information based on information such as files transmitted from 50 and past behavior information of users and organizations, and past similar information estimated by the similar information estimation unit 113 The similarity information estimation unit 113 includes a computing device including a recommendation information determination unit 114 that determines information to be recommended to the user, and the similar information estimation unit 113 performs a history based on a feature word of information such as a file transmitted from the application 150 and the current date and time. Information in the information storage unit 101 in the past under similar circumstances is searched and weighted according to the degree of coincidence, and the recommended information determination unit 114 is related from the similar information estimated by the similar information estimation unit 113. The related information in the information storage unit 103 is searched to extract information required by the application 150, and recommendation information is determined.

  Next, the configuration of information managed by each unit of the storage unit of the server 100 will be described. The history information storage unit 101 stores resource names and dates. Resources are, for example, content information such as document files and web pages, communication information such as e-mail, telephone and chat, schedule information such as meetings, visits, and work, or event information, but information with names. Any other information may be used.

  FIG. 3 is an explanatory diagram illustrating an example of a configuration of information stored in the history information storage unit 101. The history information storage unit 101 includes at least a resource ID 202 (resource ID) 201, a resource title 202, a resource type (resource type) 203, a date and time 204 related to the resource, and a feature word 205 that characterizes the resource. Are stored in association with each other. However, depending on the type of resource, data may be stored without including some information such as feature words and date / time. In addition to the information illustrated in FIG. 3, the creator or updater of the resource, the access right of the resource, and the like may be described. The date and time 204 may be stored by being divided into a creation date and time, an update date and time, a last access date and time, and the like. The resource ID 201 includes a numerical value and the like, and is automatically numbered by the history / relationship information conversion unit 111, for example.

  The person information storage unit 102 stores person information in advance. FIG. 4 is an explanatory diagram illustrating an example of a configuration of information stored in the person information storage unit 102. In the person information storage unit 102, at least a person name 302 and a person's e-mail address 303 are stored in association with a person ID (person ID) 301. The person information storage unit 102 may store accompanying information such as a belonging organization and a specialty field. The person ID is an employee number, a student number, or the like, but is a numerical value that is automatically assigned as an example when there is no appropriate number.

  The relationship information storage unit 103 stores the relationship between resources and the relationship between resources and people. FIG. 5 is an explanatory diagram illustrating an example of the configuration of the relationship information stored in the relationship information storage unit 103. The relationship information storage unit 103 stores at least a relationship source ID (relation source ID) 401, a relationship destination ID (relation destination ID) 402, and a relationship type 403 in association with each other. The relation information storage unit 103 may store a relation user name, relation weight information, relation direction information, relation date / time, update date / time, and the like. In the example illustrated in FIG. 5, R-xxxx (xxxx is a numerical value) as the related party ID 402 indicates a resource ID, and H-xxxx (xxxx is a numerical value) indicates a person ID.

  The conversion rule storage unit 104 stores a conversion rule for converting data transmitted from the application 150 into history information and related information. The conversion rule storage unit 104 stores a history information conversion rule 500 and a relationship information conversion rule 600.

  FIG. 6 is an explanatory diagram showing an example of the configuration of the history information conversion rule 500. The history information conversion rule 500 is associated with at least a data type (data type) 501, an item name 502 in the data, storage destination information 503 in the history information storage unit 101, and a pre-conversion process 504. Stored in advance. However, the data may be stored in a state where the pre-conversion processing 504 and the item name 502 in the data are not included.

  In the example illustrated in FIG. 6, three conversion rules are stored for “e-mail”. The first conversion rule described in the first row in FIG. 6 specifies that a resource ID (data identification information) is assigned to “e-mail” and stored as the resource ID 201 in the history information storage unit 101. To do. The second conversion rule described in the second row in FIG. 6 specifies that the title of “e-mail” is stored as the resource title 202 in the history information storage unit 101. The third conversion rule described in the third line in FIG. 6 specifies that the transmission date / time of “e-mail” is stored as the date / time 204 related to the resource in the history information storage unit 101.

  FIG. 7 is an explanatory diagram showing an example of the configuration of the relationship information conversion rule 600. The relation information conversion rule 600 includes at least a data type 601, an association process occurrence condition 602, an association source 603, an association destination 604, a relation type 605, a pre-conversion process 606, and history information. Registration processing information 607 is stored in advance in association with each other. However, the data may be stored in a state in which the pre-conversion process 606 and the registration process information 607 for history information are not included.

  In the example shown in FIG. 7, three conversion rules are stored for “e-mail”. In the first conversion rule described in the first row in FIG. 7, when an attached file exists in “e-mail”, the resource of “e-mail” is set as the relation source ID 401 in the relation information storage unit 103. The ID is stored, and “file ID” is stored as the related destination ID 402 in the related information storage unit 103.

Next, the function of each part of the calculation means will be described.
The history / relationship information conversion unit 111 receives data related to the resource from the application 150, searches for a conversion rule in the conversion rule storage unit 113 based on the type of data of the resource, and performs processing based on the searched conversion rule. The processing result is output to the history information storage unit 101 and the relationship information storage unit 103.

  The feature word extraction unit 112 analyzes the contents of the resources input from the history / relationship information conversion unit 111 and the similar information estimation unit 113, and returns the analysis result. As an example, the feature word extraction unit 112 has a morpheme analysis function and a feature word extraction function. When the resource is document content, the feature word extraction unit 112 analyzes the content of the document content using the morpheme analysis function and the feature word extraction function. The morphological analysis function is a function that divides an input sentence into words and gives parts of speech. As the morphological analysis function, “chasen” (URL: http://chasen-legacy.sourceforge.jp/) or “Mecab” (URL: http://mecab.sourceforge.net) can be used. Applications having other morphological analysis functions may be used.

  As a feature word extraction function, “TF / IDF method” (URL: http://en.wikipedia.org/wiki/Tf-idf and URL: http://nlp.nagaokaut.ac.jp/~sekiguti/ (See doc / pdf / tfidf / tfidf / index.html), but other methods may be used.

  The feature word extraction unit 112 narrows down a large number of feature words extracted using the feature word extraction function into feature words having a certain number of high scores. The number of cases is, for example, about 10 to 100. Next, the feature word extraction unit 112 outputs the input resource ID, the feature word obtained by the calculation, and the weight. Although it is preferable to weight the score value, this is not essential. If the score value cannot be used as a weight, the initial value (for example, 1) is used as the weight. Further, when the content is other than the document content, the feature word extraction unit 112 extracts information corresponding to the feature word by using a voice recognition function or a keyword extraction function from a moving image.

  The similar information estimation unit 113 searches information stored in the history information storage unit 101 based on the data transmitted from the application 150 and estimates similar past information. Then, the estimated past similar information is output to the recommended information determining unit 114 together with the similarity.

  The recommendation information determination unit 114 determines recommendation information that is information to be recommended (for example, a person) according to the similar information input from the similar information estimation unit 113 and the recommendation request item (recommendation request) transmitted from the application 150. Determine and send to application 150.

  Next, the application 150 will be described. For example, a mailer, a schedule management application, a document sharing application, a Web conference application, a TODO management application, a social network application, a community management application, or the like, which is used as an application 150, or software for communicating with each other. There is no software. However, these are examples of the application 150, and the application 150 is not limited thereto. The location where the application 150 exists is, for example, an arbitrary client device or server device, but does not have to exist in a specific location.

  The application 150 transmits information to the history / relationship information conversion unit 111 at an appropriate timing. The appropriate timing is the time when information is registered, updated, or newly created, such as e-mail transmission or schedule registration.

  The application 150 transmits information to the similar information estimation unit 113 at an appropriate timing. The appropriate timing is, for example, when an attached file is specified when writing an e-mail, or when a body is written when writing an e-mail. Information transmitted from the application 150 is a file, a text content such as an e-mail body, a title, a type of information (e-mail, schedule, etc.), and a type of information requested for recommendation (sender, participant, etc.). The application 150 also transmits a recommendation request item indicating the type of information to be recommended (for example, a person) when transmitting information such as the text of an electronic mail.

  The application 150 is recommended to the user based on the recommendation information received from the recommendation information determination unit 114, for example, by displaying the recommendation information on a display unit provided in a device on which the application 150 is installed. Present information.

  Next, the operation of the first embodiment will be described with reference to the flowcharts of FIGS.

  First, a process for accumulating information in the history information storage unit 101 and the relationship information storage unit 103 will be described.

  FIG. 8 is a flowchart showing the operation of the history / relationship information conversion unit 111. The operation of the history / relationship information conversion unit 111 will be described with reference to FIG. As will be described below, when the history / relationship information conversion unit 111 receives data from the application 150, the history / relationship information conversion unit 111 stores predetermined data in the relationship information storage unit 103 (see FIG. 5) based on information included in the received data. sign up.

  First, the history / relationship information conversion unit 111 receives data from the application 150 (step S701). The data is e-mail information, files, schedule information, and the like.

  Next, the history / relationship information conversion unit 111 determines the type of the received data (step S702). The method of determination is arbitrary, but for example, it is possible to determine the type of data by determining with a data format such as an extension or by registering in advance what information the application 150 transmits.

  Next, the history / relationship information conversion unit 111 searches for the history information conversion rule 500 (see FIG. 6) in the conversion rule storage unit 104 based on the determined data type (step S703). For example, when the data type is electronic mail, a row in which the column of the data type 501 in the history information conversion rule 500 is “electronic mail” is searched. The search result is used as history conversion rule information.

  Next, it is checked whether or not “feature word extraction” is present in the column of the pre-conversion processing 504 of the retrieved history conversion rule information (step S704). When there is at least one “feature word extraction”, that is, when there is a rule for performing feature word extraction processing, the history / relationship information conversion unit 111 outputs data to the feature word extraction unit 112. Then, a feature word that characterizes the resource extracted from the data input by the feature word extraction unit 112 is input from the feature word extraction unit 112 (step S705).

  Next, the history / relationship information conversion unit 111 stores information in the history information storage unit 101 based on the history conversion rule information (step S706). The history / relationship information conversion unit 111 generates an email ID according to the history conversion rule information (see FIG. 6) that specifies that an ID (resource ID) is to be given to the email, and stores it in the history information storage unit 101. Remember. Further, for example, when the history conversion rule information includes information in which the data type 501 is “e-mail”, the item name 502 in the data is “transmission date / time”, and the history storage destination 403 is “date / time”. Then, the transmission date / time information in the e-mail data is extracted, and the extracted transmission date / time information corresponding to the ID (resource ID) of the e-mail is stored in the date / time column 204 of the history information storage unit 101 (see FIG. 3). Register as Further, the feature word input from the feature word extraction unit 112 is registered as the feature word 205.

  That is, the history conversion rule information is used to set data in an appropriate column in the history information storage unit 101 based on the data transmitted from the application 150 as in the processes of steps S703 to S706. The history / relationship information conversion unit 111 assigns a new resource ID 201 when registering data in the history information storage unit 101. For example, the history information storage unit 101 automatically assigns a new resource ID 201, Resource ID 201 is assigned.

  Next, the history / relationship information conversion unit 111 searches for the relationship information conversion rule 600 (see FIG. 7) in the conversion rule storage unit 104 based on the data type (for example, e-mail) of the data received from the application 150. (Step S707). For example, if the data type 501 is “e-mail”, a row in which the column of the data type 601 in the relationship information conversion rule 600 is “e-mail” is searched (step S708). The search result is used as relation conversion rule information. A plurality of pieces of information may be searched as the relation conversion rule information.

  Next, the history / relationship information conversion unit 111 checks whether there is any data received from the application 150 that matches the relationship generation condition 602 in the relationship information conversion rule (step S709). At this time, if there is no matching condition, the related information is not stored in the related information storage unit 103 and the process is terminated. If there is a matching condition, the related information is stored in the related information storage unit 103 based on the condition (step S710). For example, when an e-mail with an attached file is sent, an item in which the data type is 601 “e-mail” and the condition 602 is “attached file” is included in the relation information conversion rule 600 shown in FIG. Therefore, the process of storing the related information in the related information storage unit 103 is performed according to the process described in the row. Specifically, in this case, an email ID is stored as the relationship source ID (relation source) 401, a file ID is stored as the relationship destination ID (relation destination) 402, and “attachment” is stored as the relationship type 403. If a person ID conversion process exists in the pre-conversion process 606, the person information storage unit 102 is searched to perform a process of converting a person's e-mail address or name into a person ID.

  The e-mail ID stored as the relation source ID 401 is an e-mail resource ID. The ID of the file stored as the related party ID 402 is an ID (resource ID) attached to the attached file. The history / relationship information conversion unit 111 attaches an ID to the attached file, for example, and stores “file” in the history information storage unit 101 as the resource type 203 together with the ID (resource ID) of the attached file.

  That is, the retrieved relationship conversion rule information associates the relationship between some pieces of information in the data transmitted from the application 150 with the items in the relationship information storage unit 103, as in the processing of steps S707 to S710. And stored in the relationship information storage unit 103.

  If the history registration process 607 (see FIG. 7) is described in the relation conversion rule information of the matched condition (step S711), the history / relationship information conversion unit 111 uses the information as a basis. The history information conversion process / related information conversion process in steps S702 to S710 is started again. For example, when an attached file exists in an e-mail, in the example shown in FIG. 7, the relation information conversion rule in which “attached file exists” in the condition 602 of the relationship information conversion rule 600 and “file” exists in the history registration process 607. Therefore, the process of storing the file in the history information storage unit 101 is started.

  That is, the data transmitted from the application 150 is repeatedly subjected to history information conversion processing / related information conversion processing according to the information therein. The process ends when there is no information to be finally registered.

  Next, processing for providing recommended information to the application 150 will be described.

  FIG. 9 is a flowchart showing the operation of the similar information estimation unit 113. The operation of the similar information estimation unit 113 will be described with reference to FIG.

  First, the similar information estimation unit 113 receives data from the application 150 (step S801). Here, the data includes character content such as a file, an e-mail body, and a title, information type (e-mail, schedule, etc.), and type of information requested for recommendation (sender, participant, etc.). Taking e-mail as an example, when the user creates an e-mail body, the application 150 transmits the e-mail body as data to the server 100, and the similar information estimation unit 113 receives the data in the server 100.

  Next, the similar information estimation unit 113 outputs a character content such as a file or an e-mail sentence in the received data to the feature word extraction unit 112, and inputs the feature word extracted by the feature word extraction unit 112 (step S802). .

  Next, the similar information estimation unit 113 searches for past similar information from the history information storage unit 101 using the feature words extracted by the feature word extraction unit 112 (step S803). For example, when the feature word of the document content is input in the process of step S802, it is assumed that there are about 10 feature words instead of one, but at least one of the feature words in the history information storage unit 101 is assumed. All rows existing in 205 are searched.

  Next, the similarity information estimation unit 113 calculates the similarity based on the retrieved data in the history information storage unit 101 (step S804). In the calculation of the similarity, the matching degree of feature words and the matching degree of data types are used. Hereinafter, an example of how to calculate the similarity will be shown.

  It is assumed that the resource transmitted from the application 150 is Q, and the information (specifically, information of the feature word 205) in the retrieved history information storage unit 101 is a1, a2, a3,. Also. The number of feature words in Q is Word_Q, the number of feature words in one piece of searched information (referred to as ak) is Word_ak, and the number of feature words that match between ak and Q is Word_ak_and_Q. Also, the degree of coincidence between the feature words of Q and ak is Word_coin_k. An example of an expression for calculating Word_coin_k is shown in Expression (1). In equation (1), the square of the number of matching feature words is divided by the number of feature words in Q and ak to obtain the matching score. However, the matching score may be obtained by other methods such as using only the matching feature word count. .

  Also, the degree of coincidence between the resource types of ak and Q is type_coin_k. An example of an equation for calculating type_coin_k is shown in equation (2). In equation (2), the degree of coincidence is changed only when the resource types match, but the similarity between resource types may be defined separately and used. Further, the value when the resource types do not match is s, but the value may be any number as long as it is an arbitrary constant smaller than 1. For example, 0.5 may be considered as the value of s.

  The final similarity between ak and Q is deg_coin_k. An example of a calculation formula for deg_coin_k is shown in Formula (3).

  The similarity is calculated for each searched information by the above formulas (1) to (3). Here, in calculating the similarity, the similarity of information is obtained from the coincidence of feature words and the coincidence of types, but other information may be used when calculating the similarity. For example, the degree of coincidence of other information such as the proximity of creation date may be further obtained, and the obtained value may be further multiplied by the result of calculation according to the expression (3). As an example, the history information storage unit 101 includes, as history information, data indicating a user (including a file creator and an e-mail sender) of data (application data) transmitted from the application 150 in the past and a use time. The similarity information determination unit 113 stores data indicating (including file creation time and e-mail transmission time), and the similar information determination unit 113 relates to information on the user and application time in the application data transmitted from the application 150 together with the recommendation request item. The similarity value to be output is changed in accordance with the similarity between the information and the data indicating the user and the data indicating the use time stored in the history information storage unit 101.

  As a specific example, the similarity information estimation unit 113 obtains the file creation time and the similarity of the usage time zone, and further multiplies the similarity obtained by the expression (3). In this case, for example, as shown in the equation (4), the larger the time, day, and month are, the larger the value is. However, elements of day of the week and year may be included.

  After calculating the similarity, the similar information estimation unit 113 sorts the similarity information by the calculated similarity (step S805), and outputs the ID and similarity of the similar information to the recommended information determination unit 114 (step S806). The process ends. At the time of output, a process for reducing output information, such as outputting only those with high similarity, may be performed.

  FIG. 10 is an explanatory diagram illustrating an example of information output to the recommended information determination unit 114. In the example illustrated in FIG. 10, the information includes ID 1001 and similarity 1002 of similar information. Similar information ID 1001 corresponds to resource ID 201 (see FIG. 3) in history information storage unit 101 in which feature word 205 is searched and the similarity of feature word 205 is high.

  FIG. 11 is a flowchart illustrating the operation of the recommendation information determination unit 114, and FIG. 12 is an explanatory diagram illustrating an example of recommendation candidate information that changes according to the processing content in the recommendation information determination unit. The operation of the recommendation information determination unit 114 will be described with reference to FIGS. 11 and 12.

  First, the recommendation information determination unit 114 inputs data from the similar information estimation unit 113 (step S901). As shown in FIG. 12A, the information to be input includes ID 1101 and similarity 1102 of similar information. Note that the data illustrated in FIG. 12A corresponds to the data illustrated in FIG. Here, the input information is referred to as similar information data.

  Next, the recommendation information determination unit 114 searches the relation information storage unit 103 (see FIG. 5) using the ID 1101 of each row in the input data, and related information (relation source ID 401 and relationship destination ID 402). ) And the relationship type 403 are acquired (step S902). When the acquired information includes person information (ID of “H-xxxx”), the recommended information determination unit 114 sets the searched person information (person ID, etc.) as related person 1103 in the similar information data. At the same time, a relationship type 1104 is added. The information added as the relationship type 1104 is information on the relationship type 403 (see FIG. 5) in the relationship information storage unit 103.

  Next, the recommendation information determination unit 114 searches for information of 2 hops ahead from the relationship information storage unit 103 (see FIG. 5) (step S903). The information of two hops ahead is information further related to information related to certain information, and can be acquired by recursively searching the relationship information storage unit 103 twice. For example, in the example illustrated in FIG. 5, when the relation source ID 401 is R-0001, the one hop destination is information of R-0003, and the two hop destination is a relation corresponding to R-0003 as the relation source ID 401. This is information such as the resource indicated by the destination ID. Similarly to the processing in step S902, if there is person information among the acquired information, the searched person information is added as related person 1103 and the relation type 1104 is added to the similar information data.

  Next, when there is a user who is operating the application 150 in the person information searched in the processes of steps S902 and S903, the recommended information determination unit 114 displays “Relationship with user 1105” in the row. The relationship information with the user is recorded in the column (step S904). Note that the recommendation information determination unit 114 identifies the user who is operating the application 150 from the data transmitted by the application 150. That is, the data transmitted by the application 150 includes information that can identify the user (for example, information indicating the sender of the e-mail). In FIG. 12B, “YES” indicates that there is a relationship with the user of the application 150. In the relationship type 1104, “sender” indicates that the user has sent an email, and “receiver” indicates that the user has received an email. That is, “Sender” is set in the relationship type 1104 and “YES” is set in the “Relationship with user 1105” column. This means that the resource (for example, e-mail) corresponding to the ID 1101 of the similar information is Indicates that 150 users were senders.

  Further, the recommendation information determination unit 114 calculates the recommendation degree weight 1106 based on the processing data (see FIG. 12B) created by the processing in steps S902 to S904 (step S905: see FIG. 12C). . Hereinafter, an example of how to calculate the recommendation weight is shown.

  The value of the “similarity” column in a specific row of input data (1 row in the table shown in FIG. 12B: bk) is bk_conf_weight, the “relationship type” column is bk_type, “ The column “Relationship with User” is bk_user_associate, and the type of data requested by the application (recommendation request item) is req_type.

  The degree of coincidence between the recommendation request item requested by the application and the relationship type in the bk data is defined as action_coin_bk. An example of the calculation formula of action_coin_bk is shown in Formula (5). The value when the relationship type does not match the recommendation request item is t, but the value when the relationship does not match may be any number as long as it is an arbitrary constant smaller than 1. For example, 0.5 may be considered as the value of t. Further, in the equation (5), the degree of coincidence is set to “1” only when the relationship type and the recommendation request item match, and the value is smaller than 1 when the relationship type does not match, but the similarity between the relationship type and the recommendation request item is similar. The degree may be defined separately and used.

  Further, the degree of relationship between bk and the user operating the application is assumed to be person_coin_bk. An example of the calculation formula of person_coin_bk is shown in Formula (6). In the equation (6), “bk_user_associate is“ YES ”” corresponds to “YES” being set in the relationship 1105 with the user in the processing data shown in FIG. The value when the user does not match is u, but the value when the user does not match is an arbitrary constant smaller than 1 as in the case of the degree of matching between the relationship type and the recommendation request item. Any number is acceptable. For example, 0.5 may be considered as the value of u. Further, in the equation (6), the degree of coincidence is changed based on the data in the “Relationship with User 1105” column in bk, but the friendship or work relationship between the users is stored separately. You may make it judge using it.

  Let the recommendation degree weight regarding bk be bk_recom_weight. An example of a formula for calculating bk_recom_weight is shown in Formula (7).

  That is, in the process of step S905, the recommendation information determination unit 114 determines the type of requested data (recommendation request item) and the degree of relationship with the user of the content stored in the relationship information storage unit 103 (specifically, the application The degree of recommendation is calculated by changing the degree of similarity based on the degree of relationship with the information transmitted by the terminal 150.

  Next, the recommendation information determination unit 114 adds the calculated recommendation degree weights for the same person (step S906: see FIG. 12D). Then, rearrangement is performed in the order of weight (step S907: see FIG. 12E). Thereafter, the result (recommendation candidate data including the recommendation candidate information illustrated in FIG. 12E) is transmitted to the application 150 (step S908), and the process ends. Here, when transmitting the results to the application 150, not all of the results may be transmitted, but the top several or several tens of results may be transmitted in order of weight.

  As described above, when processing is performed according to the flowchart of FIG. 11, the data changes as shown in FIG. 12, and recommended information (in this example, recommended person information) is transmitted to the application 150.

  The effect of the present embodiment is to analyze the contents of files and emails, search for past similar information, and recommend destination information related to the searched information, thereby responding to past similar situations in the user. It is possible to recommend an appropriate person.

  That is, in a system that recommends destination information or the like, it is possible to recommend an appropriate person according to a similar situation that the user has performed in the past. The reason for this is by analyzing the user's input files and e-mail contents, searching for communication data when dealing with similar contents in the past, and determining the person in a similar situation in the past from related information such as destination information at that time Because it does.

  In addition, another effect of this embodiment is to save communication data of a plurality of users as history information and related information in an integrated manner, and search not only the past history but also the history of others when determining recommendation information, Even if the user has never experienced a similar situation in the past by recommending a suitable person based on the past experience of other people by integrating and ranking the information of yourself and others with different weights It will be possible.

  That is, even when the user has never experienced a similar situation in the past, it is possible to recommend an appropriate person based on the past experience of other persons. The reason is that communication data of multiple users is integrated and stored as history information and related information, and when there is a recommendation request from the application, not only the past history but also other people's history are searched and recommended information By combining and ranking the information of yourself and others with different weights when scoring, maintaining the recommended function based on your past information, while also adding recommended information based on the past information of others Because it does.

Embodiment 2. FIG.
Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 13 is a block diagram showing a second embodiment of the recommendation information providing apparatus according to the present invention. In the second embodiment, a person relationship conversion unit 115 and a role information storage unit 105 are added to the configuration of the first embodiment.

  That is, the storage device is provided with a role information storage unit 105 that stores a group to which a person belongs and a role in the group, and converts a recommended person of an organization or group to which the user does not belong to the arithmetic device. A person information conversion unit 115 is provided. Then, the person information conversion unit 115 searches the role information storage unit 105 based on the person information output from the recommendation information determination unit 114, converts the person information into a role, and further recalculates the recommendation degree, Based on the above, the role information storage unit 105 is searched again to convert it to the person of the organization or group to which the user belongs, and the recommendation information is determined.

  The role information storage unit 105 stores information on groups such as organizations and groups, and role information in the groups. FIG. 14 is an explanatory diagram illustrating an example of role information stored in the role information storage unit 105. The role information stored in the role information storage unit 105 includes at least a person ID 1301, a belonging group 1302 indicating a belonging group, and a role name 1303. Basically, it is assumed that the current role name 1303 is included, but in addition, a past role history in consideration of the group enrollment period and the role charge period may be included. In the example illustrated in FIG. 14, for example, the person identified by H-0001 indicates that it is also the manager of the planning department and the leader of the A product project team.

  The person relationship conversion unit 115 once converted the data output from the recommendation information determination unit 114 into role information, converted into the corresponding person of the group to which the user using the application 150 belongs, and reconstructed the weight. After that, recommendation information is determined.

  Next, the operation of this embodiment will be described with reference to a flowchart showing the operation of the person information conversion unit 115 in FIG. 15 and an explanatory diagram showing an example of information that changes according to the processing content in FIG.

  First, the person relationship conversion unit 115 inputs recommendation information data from the recommendation information determination unit 114 (step S1401). The recommendation information data includes at least a person ID 1501 and a recommendation degree weight 1502 (see FIG. 16A). Note that the recommendation information data illustrated in FIG. 16A corresponds to the recommendation information data illustrated in FIG.

  Next, the person relationship conversion unit 115 searches the role information storage unit 105 using the person ID 1501 in the recommendation information data (step S1402). Then, role information is extracted. Generally, since a person belongs to a plurality of organizations or has a plurality of roles, even when searching for a single person ID in the role information storage unit 105, only one item is not always searched. The person relationship conversion unit 115 extracts all search results when multiple items are searched.

  Next, the person relationship conversion unit 115 adds the recommendation degree weight 1502 included in the recommendation information data to the searched role information, and adds the recommendation degree weight when there is the same role information (step S1403). ). For example, if there are two persons A and B in the recommendation information data, and the recommendation weights are 3 and 2, respectively, and both A and B have the role of “department manager”, the recommendation degree weight of “department manager” is 3 + 2 = 5.

  Next, the person relationship conversion unit 115 rearranges the role information in the order of the recommended recommendation weight (step S1404). At this point, data of a set of a role 1503 and a recommendation degree weight 1504 corresponding to the role is created (see FIG. 16B).

  Next, the person relation conversion unit 115 plays a role in the organization (for example, “planning department”) to which the user using the application 150 belongs for the role (for example, “department manager”) for which the recommendation degree weight is calculated. A search is performed from the information storage unit 105 (step S1405). Then, the search result is extracted from the role information storage unit 105. For example, when the “department manager” exists in the role list shown in FIG. 16B, the contents of the role information storage unit 105 are examined to determine whether there is a “department manager” of the organization to which the user belongs. If there are multiple search results, all are extracted.

  Next, the person relationship conversion unit 115 gives a recommendation degree weight 1504 corresponding to the role calculated in the processes up to step S1404 to the person extracted from the role information storage unit 105, and recommends a recommendation when there is the same person. The degree weights are added together (step S1406), and are rearranged in the order of the weights (step S1407). At this time, the data is a set of a person 1505 and a recommendation degree weight 1506 corresponding to the person (see FIG. 16C). Finally, the person relationship conversion unit 115 transmits the calculated person information and recommendation degree weight to the application 150 (step S1408), and ends the process.

  The recommendation information determination unit 114 basically executes the process as shown in FIG. 15, but in order to be able to recommend a more accurate person, the role charge period, the group enrollment period, and the electronic Based on the correspondence relationship with the mail transmission date and time, the roles at the past time may be taken into consideration, and the past roles may be weighted and summed up for the person concerned.

  The effect of this embodiment is that the recommended person is once converted into a role, and based on that role, converted into a person of a group or organization to which the user belongs and re-ranked, so that another department or It is possible to recommend an appropriate person from his / her department or organization based on the history of another organization.

  In other words, when recommending an appropriate person to a user based on the experience of another person, recommending an appropriate person from his / her department or organization based on the history of other departments or organizations Is possible. The reason is that after estimating the similar situation and determining a recommended person, the person is once converted into a role, and then converted into a person of a group or organization to which the user belongs based on the role and reranked.

  In each of the above embodiments, data determination means (history / relation information conversion unit 111) for determining the data type (for example, e-mail, file) of application data transmitted from a processing device that executes processing according to application software. The history information storage means stores the data type determined by the data determination means in association with the data identification information, and the similar information determination means determines the recommendation request item from the processing device by the data determination means. Also included is a form in which the similarity value to be output is changed according to the data type similarity that is the similarity between the data type determined for the application data transmitted together with the data type stored in the history information storage means. ing.

  Each of the above embodiments is realized by the recommended information candidate storage unit (person information storage unit 102) that stores information to be recommended information in advance in association with information data identification information (for example, a person ID). The relationship information storage means stores the data identification information of the application data transmitted from the application in the past or the information data identification information stored in the recommended information candidate storage means as the relationship destination information. include.

  In each of the above-described embodiments, the history information storage means includes data indicating application data users (including e-mail creators and file users) transmitted from the application in the past as history information. And data indicating the use time (e-mail transmission date / time, file creation date / time, file use date / time), and information related to the user in the application data transmitted by the similar information determining means together with the recommendation request item from the processing device. There is also included a form in which the similarity value to be output is changed in accordance with the similarity between the information related to the time and the data indicating the user and the data indicating the use time stored in the history information storage means.

  Next, a specific example will be described with reference to the screen examples of FIGS. FIG. 17 is an explanatory diagram illustrating a screen example of an email creation screen 1600 by a mailer as an example of the application 150. It is assumed that the e-mail creation screen 1600 includes at least a destination input field 1601, a subject name input field 1602, a main body text field 1603, and an attached file designation field 1604.

  FIG. 18 is an explanatory diagram illustrating an example of a person information recommendation screen. The person information recommendation screen is automatically presented by the processing of the application 150 while the application 150 is being used, or is presented by the application 150 when an arbitrary operation is performed by the user while the application 150 is operating.

  For example, when an e-mail creation screen as shown in FIG. 17 is displayed, if the user inputs a text or an attached file, the application 150 automatically analyzes the content and the person information recommendation screen 1701. Or is displayed when the application 150 clicks the destination information setting button 1605. The application 150 inputs information indicating the person selected by the user on the person information recommendation screen 1701. For example, when the application 150 is a mailer, the selected person enters the destination column.

  When the user uses a mailer as the application 150, when the user inputs a text or an attached file, the application 150 transmits the text information and the attached file to the server. In the server, after the feature word extraction unit 112 shown in FIG. 2 extracts the feature words from the text information and the attached file, the similar information estimation unit 113 uses the extracted feature words as a past similar e-mail. And the history information storage unit 101 are searched. Then, the destination information when the electronic mail or file is transmitted is searched from the related information storage unit 103, and the recommendation information is determined. Thereafter, a person list based on the recommendation information is transmitted to the application 150, and the application 150 creates a person information recommendation screen.

  While creating an email using the email creation screen by the mailer as shown in FIG. 17, the user inputs a destination using the person information recommendation screen 1701 as shown in FIG. 18 and transmits the email. . The transmitted e-mail is output to the history / relationship information conversion unit 111, and is used when a user or another user transmits information later.

  FIG. 19 is an explanatory diagram showing a screen example in another embodiment. In the example shown in FIG. 19, when there is a document file 1801 on the user's client device (for example, personal computer), when the user right-clicks or mouses over the document file, the application 150 displays an action candidate window 1802. indicate. For example, when the user selects “e-mail”, the application 150 requests the server for a person recommendation calculation according to the selection action, and transmits the request from the recommendation information determination unit 114 shown in FIG. The received person list is received and displayed on the person information recommendation screen 1803. When the user selects a person from the displayed person list, the e-mail creation screen 1600 is launched with the person in the destination and the selected document file registered in the attached file.

  When the application 150 and the server perform such an operation, the user can write an e-mail using the recommended information without complicated operations.

  Further, the example shown in FIG. 19 can be applied to telephone, chat, Web conference, new schedule registration, etc. in addition to electronic mail. When applied to other than e-mail, the basic operation is the same as that of e-mail, but the target behavior is different, and therefore the person information transmitted from the recommended information determination unit 114 is partially different.

  As an example of the situation of receiving the recommendation, in addition to the example of the e-mail creation screen as shown in FIG. 17 and the example of the file management in the user client device as shown in FIG. It can also be used for information browsing during file reading (when it is desired to send a reply e-mail, related e-mail, etc. during reading) or for creating a communication partner during file creation.

  When dealing with some document contents, recommend requests in all situations related to document contents such as telephone, e-mail, chat, web conference, schedule registration, etc. Is possible.

  INDUSTRIAL APPLICABILITY The present invention is applicable to a system that collaborates with a plurality of people, such as a mailer, a web conference management system, a schedule management system, a community management system, a chat (instant messenger) system, a groupware, a social network management system, and a project management system. There is but is not limited to this. As another example, the present invention can be applied to uses such as a knowledge management system, a blog management system, and a document management system.

It is a block diagram which shows the outline | summary of this invention. It is a block diagram which shows the structure of the 1st Embodiment of this invention. It is explanatory drawing which shows the example of a structure of the information memorize | stored in a log | history information storage part. It is explanatory drawing which shows the example of a structure of the information memorize | stored in a person information storage part. It is explanatory drawing which shows the example of a structure of the relationship information memorize | stored in a relationship information storage part. It is explanatory drawing which shows the example of a structure of a log | history information conversion rule. It is explanatory drawing which shows the example of a structure of a relationship information conversion rule. It is a flowchart which shows operation | movement of the log | history and related information conversion part in 1st Embodiment. It is a flowchart which shows operation | movement of the similar information estimation part in 1st Embodiment. It is explanatory drawing which shows the example of the information output to a recommendation information determination part. It is a flowchart which shows operation | movement of the recommendation information determination part. It is explanatory drawing which shows the example of the recommendation candidate information which changes according to the processing content in a recommendation information determination part. It is a block diagram which shows the structure of the 2nd Embodiment of this invention. It is explanatory drawing which shows the example of the role information memorize | stored in a role information storage part. It is a flowchart which shows operation | movement of the person information conversion part in 2nd Embodiment. It is explanatory drawing which shows the example of the information which changes according to the processing content of the person information conversion part in 2nd Embodiment. It is a figure which shows the example of an e-mail preparation screen by an application. It is a figure which shows the example of a person information recommendation screen by an application. It is explanatory drawing which shows the example of a screen after a user selects content using an application until a communication candidate person is recommended.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Relation information storage means 2 History information storage means 3 Similar information determination means 4 Recommendation information determination means 5 Processing apparatus 100 Server 101 History information storage part 102 Person information storage part 103 Relation information storage part 104 Conversion rule storage part 105 Role information storage part 111 History / Relation Information Conversion Unit 112 Feature Word Extraction Unit 113 Similar Information Estimation Unit 114 Recommended Information Determination Unit 115 Person Information Conversion Unit

Claims (15)

A recommended information providing device that provides recommended information to a user in order to support information input by the user,
History information that stores, as history information, characteristic data that characterizes the application data extracted from application data transmitted from a processing device that executes processing according to application software, in association with data identification information attached to the application data Storage means;
As the relationship information, the data identification information of the application data transmitted from the processing device is used as the relationship source information, the information related to the application data is stored as the relationship destination information, and the relationship type indicating the relationship between the relationship source and the relationship destination Related information storage means for storing
Data identification information corresponding to feature data having a relatively high similarity by calculating the similarity between the application data transmitted together with the recommendation request item from the processing device and the feature data stored in the history information storage means A similarity information determining means for outputting the
Based on the data identification information output by the similar information determination means, the storage contents of the relation information storage means are searched, the relation type corresponding to the relation destination information that matches the recommendation request item, and the user of the application of the relation type And a recommendation information determining means that uses the value obtained by changing the similarity as a recommendation level of the related party information and uses the related party information having a relatively high recommendation level as recommended information. A recommendation information providing device characterized by the above. As history information conversion rules, history information conversion rule storage means in which items to be extracted as history information from application data transmitted from the processing device are set,
The relationship information conversion rule includes a relationship information conversion rule storage unit in which the type of relationship source information, the type of relationship destination information, and the relationship type are set according to the type of application data,
The history information storage means stores, as history information, data obtained by converting application data transmitted from the processing device based on the history information conversion rule,
The recommendation information providing apparatus according to claim 1, wherein the relation information storage unit stores data obtained by converting application data transmitted from the processing apparatus based on the relation information conversion rule as relation information. A feature word extraction means for extracting feature words from the data;
The history information storage means stores the feature words extracted by the feature word extraction means as feature data,
The similarity information determination unit outputs the similarity between the feature word extracted by the feature word extraction unit with respect to the application data transmitted together with the recommendation request item from the processing device and the feature word stored in the history information storage unit. It is set as similarity. The recommendation information provision apparatus of Claim 1 or Claim 2. Comprising a data discriminating means for discriminating the data type of application data transmitted from a processing device that executes processing according to application software;
The history information storage means stores the data type determined by the data determination means in association with the data identification information,
The similarity information determination means is a data type similarity that is a similarity between the data type determined by the data determination means for the application data transmitted together with the recommendation request item from the processing device and the data type stored in the history information storage means. The recommendation information providing apparatus according to claim 3, wherein the similarity value to be output is changed according to the degree. Comprising recommended information candidate storage means for storing in advance information to be recommended information in association with information data identification information;
The relationship information storage means stores, as related information, data identification information of application data transmitted from an application in the past or information data identification information stored in the recommended information candidate storage means. The recommendation information provision apparatus of any one of these. The recommended information providing apparatus according to claim 5, wherein the information data identification information is information indicating a person. The history information storage means stores, as history information, data indicating a user of application data transmitted from an application in the past and data indicating a use time,
The similar information determining means indicates information on the user and information on the use time in the application data transmitted together with the recommendation request item from the processing device, and indicates data and use time stored in the history information storage means. The recommendation information providing apparatus according to any one of claims 1 to 6, wherein a value of the similarity to be output is changed in accordance with the similarity to the data. The recommendation information determination means outputs the person identification information as related party information,
Role information storage means for storing in advance group identification information indicating a group to which the person belongs and role information indicating the role of the person in association with the person identification information indicating the person;
The recommendation information determined by the recommendation information determination unit is input as data of the recommendation level and the person identification information, the input data is converted into the recommendation level of role information, and the recommendation level output by the recommendation information determination unit is used in the application. The recommendation information providing apparatus according to any one of claims 1 to 7, further comprising person relationship conversion means for converting and outputting the recommendation degree of role information in a group to which a user belongs. The person relationship conversion means inputs the recommendation degree determined by the recommendation information determination means and the person identification information, assigns the recommendation degree to the role information corresponding to the person identification information retrieved from the role information storage means, and assigns the recommendation. The degree of recommendation recommendation is added to the same role information, the person information identification in the group to which the user of the application data belongs is retrieved from the role information storage means, and the degree of recommendation recommendation corresponding to the retrieved person identification information The recommendation information providing apparatus according to claim 8. A recommended information providing method for providing recommended information to a user in order to support information input by the user,
As history information, feature data characterizing the application data extracted from application data transmitted from a processing device that performs processing according to application software is associated with data identification information attached to the application data, and history information storage means Processing to be stored in
As the relationship information, the data identification information of the application data transmitted from the processing device is used as the relationship source information, the information related to the application data is stored as the relationship destination information, and the relationship type indicating the relationship between the relationship source and the relationship destination Storing the information in the related information storage means;
Data identification information corresponding to feature data having a relatively high similarity by calculating the similarity between the application data transmitted together with the recommendation request item from the processing device and the feature data stored in the history information storage means To output with the similarity,
Based on the output data identification information, the stored contents of the relation information storage means are searched, and the relation type corresponding to the relation destination information that matches the recommendation request item and the relation with the user of the application of the relation type A recommended information providing method comprising: performing a process of setting a value obtained by changing the similarity as a recommendation level of related party information and using related party information having a relatively high recommendation level as recommended information. . A recommendation information providing method executed by an apparatus having role information storage means for storing in advance group identification information indicating a group to which a person belongs and role information indicating a role of the person in association with person identification information indicating a person. And
In the process of using the related information with a relatively high recommendation level as the recommended information, the person identification information is output as the related information,
The data of the recommendation level and the person identification information output in the process using the related party information having a relatively high recommendation level as recommendation information is input, the input data is converted into the recommendation level of role information, and the recommendation level is The recommendation according to claim 10, wherein the recommendation degree determined by the process using the relatively high related party information as recommendation information is converted into the recommendation degree of role information in a group to which a user who is using the application belongs and output. Information provision method. A recommendation information providing system including a server device that provides recommended information to a user to support user information input, and a processing device that executes processing according to application software,
The server
History information storage means for storing characteristic data characterizing the application data extracted from the application data transmitted from the processing device as history information in association with data identification information attached to the application data;
As the relationship information, the data identification information of the application data transmitted from the processing device is used as the relationship source information, the information related to the application data is stored as the relationship destination information, and the relationship type indicating the relationship between the relationship source and the relationship destination Related information storage means for storing
Data identification information corresponding to feature data having a relatively high similarity by calculating the similarity between the application data transmitted together with the recommendation request item from the processing device and the feature data stored in the history information storage means A similarity information determining means for outputting the
Based on the data identification information output by the similar information determination means, the storage contents of the relation information storage means are searched, the relation type corresponding to the relation destination information that matches the recommendation request item, and the user of the application of the relation type A recommendation information determining means that uses a value obtained by changing the degree of similarity as a recommendation level of related party information and recommends related party information having a relatively high recommendation level as recommendation information. A recommended information provision system. The recommendation information determination means outputs the person identification information as related party information,
Server device
Role information storage means for storing in advance group identification information indicating a group to which the person belongs and role information indicating the role of the person in association with the person identification information indicating the person;
The recommendation information determined by the recommendation information determination unit is input with the data of the recommendation level and the person identification information, the input data is converted into the recommendation level of the role information, and the recommendation level output by the recommendation information determination unit is being used by the application. The recommendation information providing system according to claim 12, further comprising person relationship conversion means for converting and outputting the recommendation degree of role information in a group to which a user belongs. A recommendation information providing program installed in a computer for providing recommended information to a user to support user information input,
On the computer,
As history information, feature data characterizing the application data extracted from application data transmitted from a processing device that performs processing according to application software is associated with data identification information attached to the application data, and history information storage means Processing to be stored in
As the relationship information, the data identification information of the application data transmitted from the processing device is used as the relationship source information, the information related to the application data is stored as the relationship destination information, and the relationship type indicating the relationship between the relationship source and the relationship destination Storing the information in the related information storage means;
Data identification information corresponding to feature data having a relatively high similarity by calculating the similarity between the application data transmitted together with the recommendation request item from the processing device and the feature data stored in the history information storage means To output with the similarity,
Based on the output data identification information, the stored contents of the relation information storage means are searched, and the relation type corresponding to the relation destination information that matches the recommendation request item and the relation with the user of the application of the relation type A recommendation information providing program for executing a process in which a value obtained by changing the similarity is used as a recommendation level of related party information, and related party information having a relatively high recommendation level is used as recommendation information. Corresponding to the person identification information indicating a person, a computer provided with role information storage means for storing in advance group identification information indicating a group to which the person belongs and role information indicating the role of the person,
In the process of using the related information with relatively high recommendation level as the recommended information, the person identification information is output as the related information,
The data of the recommendation level and the person identification information output in the process using the related party information having a relatively high recommendation level as recommendation information is input, the input data is converted into the recommendation level of role information, and the recommendation level is 15. The recommendation according to claim 14, wherein the recommendation level determined in the process of using relatively high related party information as recommendation information is converted into a recommendation level of role information in a group to which a user using the application belongs and output. Information provision program.

Images