JP2010198492A - Method, device, and program for estimating information selection index, and method and system for providing information - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To prevent the reduction of a satisfaction degree in a user with respect to information to be provided. <P>SOLUTION: A correspondence table storage part 14a stores known data including a domain, a category, and a purpose definitude degree. A data interpolating part 14b establishes a correspondence table by interpolating unknown data through the use of the correspondence table of the same user or that of a similar user when the unknown data exists in the stored correspondence table. An estimating part 14c estimates the domain and category of the user at the present time, based on a behavior log stored in a behavior log storage part 12. A purpose definitude degree acquiring part 14d obtains the purpose definitude degree by searching the correspondence table by using the domain and category estimated by the estimating part 14c. An information selecting part 15 selects the information, based on the purpose definitude degree obtained by the purpose definitude degree acquiring part 14d, and provides the information to an applicable user terminal. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an information selection index estimation method, an information selection index estimation device, an information selection index estimation program, an information provision method, and an information provision system.

  Conventionally, the amount of information provided in communication networks such as the Internet has been increasing explosively. For this reason, it has become difficult for users who use communication networks to actively search for their intended information, and in recent years, information providing systems that provide information suitable for individual users have been developed. It is.

  In such an information providing system, information suitable for individual users is provided using the following two methods.

  The first method is a method for estimating the purpose of the user's action and providing information that matches the estimated purpose. For example, in the first method, the scheduled action that the user may possibly take is estimated according to the situation of the user, and the purpose of the scheduled action is estimated at the timing that the estimated scheduled action may occur. (For example, see Patent Document 1).

  The second method specifies an attribute such as a user's interest and interest, and provides information according to the specified attribute. For example, in the second method, an action model (user attribute) is constructed from an operation history (action history) in a user terminal owned by the user, and information of interest to the user is provided according to the action model ( For example, see Patent Document 2).

JP 2005-71026 A JP 2000-99486 A

  By the way, the above-described conventional technique has a problem that the user's satisfaction may not be obtained with respect to the provided information.

  That is, the above-described conventional first method provides information that matches the purpose of the user. However, if the user is looking for information on various topics, not just the topic that the user is currently watching, the information provided by the first method is the topic currently being watched, The user cannot be alerted and the user's satisfaction cannot be obtained.

  In addition, the above-described second conventional method provides information that matches user attributes such as interest. However, if you ’re looking for information that ’s related to the topic you ’re currently looking at, and you ’re not looking for irrelevant information, the second method is not just information related to the topic you ’re looking at, Since irrelevant information is also provided, user satisfaction cannot be obtained.

  Accordingly, the present invention has been made to solve the above-described problems of the prior art, and an information selection index estimation method capable of preventing the user's satisfaction with respect to provided information from being lowered, An object is to provide an information selection index estimation device, an information selection index estimation program, an information provision method, and an information provision system.

  In order to solve the above-described problems and achieve the object, this method is, in one aspect, an information selection index estimation method for estimating an information selection index used for selecting information desired by a user, Based on the behavior history storage step of storing the behavior history of the user in the behavior history storage unit, the behavior history stored in the behavior history storage unit, the information intended by the user at the present time, and the user at the present time It is a requirement to include an acquisition step of acquiring, as the information selection index, a purpose clarity that is a degree indicating a dependency relationship with the topic being watched.

  Further, in one aspect, this apparatus is an information selection index estimation apparatus for estimating an information selection index used for selecting information desired by a user, and an action history storage unit that stores the user's action history And a degree of object clarity, which is a degree indicating a dependency relationship between the information intended by the user at the present time and the topic that the user is currently looking at based on the behavior history stored in the behavior history storage unit As an information selection index.

  Further, in one aspect, this program is an information selection index estimation program for causing a computer to execute an information selection index estimation method for estimating an information selection index used for selecting information desired by the user, Based on the action history storage procedure for storing the action history in the action history storage unit, the action history stored in the action history storage unit, and the user's current target information, It is a requirement to cause a computer to execute an acquisition procedure for acquiring a purpose clarity, which is a degree indicating a dependency relationship with a topic, as the information selection index.

  According to another aspect of the present invention, in another aspect, the information providing apparatus includes: a user terminal operated by a user; and an information providing apparatus that selects and provides information desired by the user to the user terminal. An information providing method to be applied, the behavior history storing step of storing the user's behavior history in the behavior history storage unit, and the behavior history stored in the behavior history storage unit, the user is currently aiming And an acquisition step of acquiring a purpose clarity level that is a degree indicating a dependency relationship between the information and the topic that the user is currently paying attention to, based on the objective clarity level acquired by the acquisition step, And an information selection step of selecting information to be provided to the user terminal.

  Further, in one aspect, this system is an information providing system configured by a user terminal operated by a user and an information providing apparatus that selects and provides the user terminal with information desired by the user, The information providing apparatus includes: an action history storage unit that stores the user's action history; information that the user is currently interested in based on the action history stored in the action history storage unit; An acquisition unit that acquires a degree of clarity of purpose, which is a degree indicating a dependency relationship with the topic being watched at, and selects information to be provided to the user terminal based on the level of clarity of purpose acquired by the acquisition unit And information selection means.

  According to the disclosed method, apparatus, program, and system, it is possible to prevent the user's satisfaction with the provided information from decreasing.

FIG. 1 is a diagram for explaining the overall configuration of an information providing system including the information providing apparatus according to the first embodiment. FIG. 2 is a block diagram illustrating a configuration of the information providing apparatus according to the first embodiment. FIG. 3 is a diagram (1) for explaining the action log. FIG. 4 is a diagram (2) for explaining the action log. FIG. 5 is a block diagram illustrating a configuration of the purpose clarity degree estimation unit according to the first embodiment. FIG. 6 is a diagram for explaining the correspondence table storage unit. FIG. 7 is a diagram for explaining the data interpolation unit. FIG. 8 is a flowchart for explaining the correspondence table determination process of the information providing apparatus according to the first embodiment. FIG. 9 is a flowchart for explaining information providing processing of the information providing apparatus according to the first embodiment. FIG. 10 is a block diagram illustrating a configuration of the purpose clarity degree estimation unit according to the second embodiment. FIG. 11 is a flowchart for explaining information providing processing of the information providing apparatus according to the second embodiment. FIG. 12 is a block diagram illustrating a configuration of a purpose clarity degree estimation unit according to the third embodiment. FIG. 13 is a flowchart for explaining information providing processing of the information providing apparatus according to the third embodiment.

  Exemplary embodiments of an information selection index estimation method, an information selection index estimation device, an information selection index estimation program, an information provision method, and an information provision system according to the present invention will be described below in detail with reference to the accompanying drawings. Hereinafter, a case where an information providing apparatus that executes the information selection index estimation method according to the present invention is incorporated in an information providing system will be described as an example.

  First, the overall configuration of the information providing system including the information providing apparatus according to the first embodiment will be described. FIG. 1 is a diagram for explaining the overall configuration of an information providing system including the information providing apparatus according to the first embodiment.

  As illustrated in FIG. 1, the information providing apparatus 10 according to the first embodiment is connected to each user terminal of the user terminal group 20 and each content server of the content server group 30.

  The user terminal group 20 includes a plurality of user terminals (for example, user terminals (1) to (n) shown in FIG. 1). The user terminal includes, for example, a personal computer, a PDA, or a device having a communication function such as a mobile phone or a PHS, and includes a dedicated application for using the network, such as a Web browser. The user terminal may be a device having a GPS function.

  The content server group 30 holds content such as text and images, and in response to a distribution request from a user terminal, a plurality of content servers (for example, the content server (1) to content shown in FIG. 1). Server (m)). In addition, each user terminal which comprises the user terminal group 20 and each content server which comprises the content server group 30 are each connected by the network network which is not shown in figure, The content server which received the delivery request | requirement of the content from the user terminal Distributes content corresponding to the distribution request to the user terminal that has made the distribution request.

  The information providing apparatus 10 selects information desired by the user (for example, information on a URL for requesting distribution of the content desired by the user to the content server) and provides the selected information to the user terminal operated by the user. .

  Here, the information providing apparatus 10 according to the first embodiment selects information using the purpose clarity described in detail below and provides the information to the user terminal, so that user satisfaction with the provided information is reduced. The main feature is that this can be prevented.

  This main feature will be described with reference to FIGS. FIG. 2 is a block diagram illustrating the configuration of the information providing apparatus according to the first embodiment. FIGS. 3 and 4 are diagrams for explaining the action log. FIG. 5 illustrates the purpose clarity estimation according to the first embodiment. FIG. 6 is a diagram for explaining the correspondence table storage unit, and FIG. 7 is a diagram for explaining the data interpolation unit.

  As illustrated in FIG. 2, the information providing apparatus 10 according to the first embodiment includes an action log acquisition unit 11, an action log storage unit 12, a content acquisition unit 13, a purpose clarity degree estimation unit 14, and an information selection unit 15. Have

  The action log acquisition unit 11 acquires an action log of each user for which the information providing apparatus 10 provides information from each user terminal constituting the user terminal group 20. Then, the action log acquisition unit 11 stores the acquired action log in the action log storage unit 12 in association with the ID (user ID) of the user who operates the user terminal that is the acquisition destination.

  The behavior log storage unit 12 stores a behavior log for each user acquired by the behavior log acquisition unit 11.

  Here, the content of the action log acquired by the action log acquisition unit 11 and stored in the action log storage unit 12 will be described with reference to FIG. As shown in FIG. 3, the behavior log includes “user terminal operation / use history”, “user current location and movement history”, “user product purchase history”, and “user schedule information”. And “the surrounding situation surrounding the user”.

  As specific examples of “operation / usage history of user terminal”, as shown in FIG. 3, “history of web pages viewed”, “history of keywords input to search engine”, “started or Examples include “history of activated applications”, “window titles of activated applications”, “operation history of input devices such as a mouse and a keyboard”, and “transmission / reception history of electronic mail”. As a method for obtaining the “operation / use history of user terminal”, as shown in FIG. 3, “recording by a recording application” provided in the user terminal can be cited.

  As specific examples of the “user's current position and movement history”, as shown in FIG. 3, “position expressed by latitude and longitude”, “description of place by tag (eg, home, work, station, Movie theaters, department stores, amusement parks) ”. As a method for obtaining “user's current position and movement history”, as shown in FIG. 3, “GPS positioning” provided in the user terminal can be cited.

  In addition, as a method for acquiring the “history of product purchase of the user”, as shown in FIG. As a method for obtaining “user schedule information”, as shown in FIG. 3, information collection from “groupware” provided in the user terminal can be mentioned.

  Further, as a specific example of “a surrounding situation surrounding the user”, as shown in FIG. 3, “time, weather, temperature, accompanying person” and the like can be cited. As a method of acquiring “the surroundings surrounding the user”, as shown in FIG. 3, “observation by an observation function” provided in the user terminal can be cited.

  Here, in the following, the action log acquisition unit 11 performs “history of browsed web pages” that is “operation / use history of user terminal” and “location of tag location” that is “user's current location and movement history”. A case where “description” is acquired from a user terminal having a GPS function and stored in the action log storage unit 12 will be described. Note that the present invention can be realized even when any of the various behavior logs described with reference to FIG. 3 is stored in the behavior log storage unit 12.

  Specifically, the action log storage unit 12 in this embodiment includes a “user ID” for uniquely identifying a user, a “start time” when the user starts browsing the web page, and the user The “end time” when the browsing of the web page is terminated, the “URL” input to the user terminal by the user to browse the web page, and the current location of the user acquired from the GPS function provided in the user terminal “ An action log associated with “GPS information” is stored.

  For example, as shown in FIG. 4A, the action log storage unit 12 includes “user ID: 001”, “start time: 10/15 12:20”, and “end time: 10/15 12: 25 ”,“ URL: http://aaa.co.jp/blog ”, and“ GPS information: P1 ”in association with each other are stored.

  The behavior log stored in the behavior log storage unit 12 may be stored separately for each user (for each user ID) as shown in FIG.

  Returning to FIG. 2, the content acquisition unit 13 obtains the corresponding content from each content server constituting the content server group 30 based on the content acquisition request from the purpose clarity estimation unit 14 described later and the information selection unit 15 described later. get. The content acquisition unit 13 will be described in detail later.

  Based on the behavior log stored in the behavior log storage unit 12, the purpose clarity degree estimation unit 14 provides information on a user's current objective for each user who operates the user terminals constituting the user terminal group 20, the user Estimate the purpose clarity, which is the degree of dependence on the topic that is currently being watched.

  Here, “the purpose clarity is high” means that the information requested by the user at the present time is “information related to the topic currently being watched”, and “ It is a situation that is not “unrelated information”. That is, since the information estimated by the user who is estimated as “high degree of object clarity” is clearly determined, the information providing apparatus 10 is narrow with respect to the user regarding the “topic currently being watched”. What is necessary is just to show the range information.

  On the other hand, “the purpose clarity is low” is a situation that does not depend on whether or not the information requested by the user at the present time is related to “the topic being watched at the present time”. That is, since the information estimated by the user who is estimated as “the purpose clearness is low” has not been determined, the information providing apparatus 10 has, for example, “ Information according to attributes such as the user's interests and interests may be presented with priority.

  Therefore, in order to estimate the purpose clarity, the goal clarity estimation unit 14 in the first embodiment, as shown in FIG. 5, includes a correspondence table storage unit 14a, a data interpolation unit 14b, an estimation unit 14c, and a goal clarity. 14d.

  The correspondence table storage unit 14a stores a correspondence table in which the domain at the time when the user's action is executed, the genre of the content of the user's action, and the above-described purpose clarity are associated with each other. In the present embodiment, a case where the object clarity is expressed as a real value from “0.0 to 1.0” will be described.

  Here, the “domain” is a classification indicating what “position” the user is acting in, for example, work, meal, hobby, housework, travel, and the like.

  The “genre” is a keyword representing the field to which the action content belongs, and examples thereof include academic, economic, daily life, education, culture, sports, and entertainment.

  For example, the administrator of the information providing apparatus 10 collects the purpose clarity in the combination of domain and genre from each user in advance through a questionnaire survey or the like, and stores the collected information (known data) in the correspondence table storage unit 14a. Store as.

  Thereby, for example, as shown in FIG. 6, the correspondence table storage unit 14 a has a target clarity of “0” when the user of “user ID: 001” executes “genre: c” in “domain: A”. .2 "is stored.

  Note that the input process of the correspondence table may be a case where it is manually input by the administrator as described above, or the information providing apparatus 10 transmits a questionnaire to each user terminal, and the completed questionnaire is displayed. It may be a case where the known data is automatically stored in the correspondence table storage unit 14a as a correspondence table after being received from each user terminal.

  Returning to FIG. 5, when the correspondence table is not completed with the known data collected from the user, the data interpolation unit 14 b interpolates unknown data that is indeterminate by machine learning. That is, the data interpolation unit 14b estimates the unknown data and determines the correspondence table by using the known data as the teacher data. Here, the data interpolation unit 14b interpolates unknown data by two interpolation methods described below, and determines the correspondence table stored in the correspondence table storage unit 14a.

  The first interpolation method is a method of interpolating unknown data using known data of a domain similar to the domain to be interpolated in the correspondence table of the same user. For example, as shown in FIG. 7A, the data interpolation unit 14b has a clear purpose of the unknown data “domain: A, genre: a” in the correspondence table of “user ID: 001”. When the degree is interpolated, “domain: C, genre: a, purpose clarity: 0.1”, which is known data of the domain “C” similar to the domain “A”, is extracted.

  Then, the data interpolation unit 14b performs the “object clarity: 0.1” of the known data based on the similarity between “domain: A, genre: a” and “domain: C, genre: a”. Weighting is performed to calculate the purpose clarity of “domain: A, genre: a”.

  The second interpolation method is a method of interpolating unknown data using known data in a similar user correspondence table. For example, as illustrated in FIG. 7B, the data interpolation unit 14 b calculates the similarity between the correspondence table of “user ID: 001” and the correspondence table of other users, and the correspondence of “user ID: 001”. As the correspondence table having the highest similarity to the table, the correspondence table of “user ID: 002” is extracted.

  Then, based on the similarity between the correspondence tables, for example, the data interpolation unit 14b weights the object clarity “0.2” in “domain: D, genre: b” of “user ID: 002”. Then, the purpose clarity in “domain: D, genre: b” of “user ID: 001” is calculated.

  Returning to FIG. 5, the estimation unit 14 c estimates the user domain and genre based on the behavior log stored in the behavior log storage unit 12. Specifically, the estimation unit 14c estimates the current domain and genre of the user based on the action log stored in the action log storage unit 12.

  Specifically, the estimation unit 14c acquires the user's position from “GPS information” in the action log illustrated in FIG. 4A or FIG. 4B, and from the acquired user position, Estimate the current domain. For example, the estimation unit 14c stores the correspondence relationship between the user's position and the domain in advance, and searches the domain corresponding to the acquired user's position to estimate the current domain of the user.

  Note that the current domain of the user may be obtained by the information providing apparatus 10 when the user manually inputs the user terminal to the user terminal. Further, in the present embodiment, the case where the current position of the user is acquired by the “GPS function” provided in the user terminal has been described. However, the present invention is not limited to this, for example, the “IP address of the user terminal The current position of the user may be acquired from “”.

  Furthermore, the estimation unit 14c requests the content acquisition unit 13 to acquire the text of the web page “URL” in the action log illustrated in FIG. 4A or 4B. And the estimation part 14c estimates a user's present genre by extracting the feature word from the text of the web page which the content acquisition part 13 acquired.

  Here, the estimation unit 14c is, for example, ““ Observation platform system that discovers a keyword that represents the feature of a sentence and extracts an important sentence ”, IEICE Transactions D-1, vol. J84-D-1, No. 2, pp. 146-154, February 2001 ”, feature words are extracted, and the user's current genre is estimated.

  The purpose clarity level acquisition unit 14d acquires the goal clarity level corresponding to the current domain and genre of the user estimated by the estimation unit 14c by searching the correspondence table stored in the correspondence table storage unit 14a. For example, if the estimated “domain and genre” of the user who is “user ID: 002” at the present time is “D and c” by the estimating unit 14 c, the purpose clarity obtaining unit 14 d in FIG. The correspondence table of “user ID: 002” shown in B) is searched to obtain “objective clarity: 0.8”.

  Here, as the time when the estimation unit 14c estimates the domain and the genre and the purpose clarity level acquisition unit 14d acquires the goal clarity level (time of goal clarity estimation time), the timing described below can be cited.

  The first time is a time when a new action log is stored in the action log storage unit 12. That is, each time a user newly browses a web page and a new action log is stored, the estimation unit 14c estimates the genre and domain, and the purpose clarity level acquisition unit 14d determines the goal clarity level. get.

  The second time is when the “GPS information” in the action log storage unit 12 has changed. That is, every time the current position of the user changes, the estimation unit 14c estimates the genre and the domain, and the purpose clarity level acquisition unit 14d acquires the goal clarity level.

  The third time is when the user's “request” changes. Here, the “request” means “motivation for user behavior at a certain point in time”.

  Specifically, the estimation unit 14c analyzes the web page text corresponding to the “URL” of the action log, and based on the similarity between the web page texts, the browsing content of the user stored in the currently stored action log is obtained. Clustering. Thereby, the estimation unit 14c clusters a series of action logs of the same user for each “request”. For example, the estimation unit 14c may be configured as “" Web browsing history classification method focusing on user request change ”, Information Processing Society of Japan, 2008-NL-187 (10), 2008/9/25, pp.65. The browsing content clustering process is executed by the method described in “−70”.

  Then, every time a new action log is stored in the action log storage unit 12, the estimation unit 14c analyzes the web page text corresponding to the “URL” of the action log, and the analysis result of the new web page text When the current browsing content has changed from the browsing content until just before, it is determined that the “request”, which is a motivation for the user's behavior, has changed. Then, every time the “request” changes, the estimation unit 14c estimates the genre and the domain, and the purpose clarity level acquisition unit 14d acquires the goal clarity level.

  When the current domain is acquired by the user's input, the estimation unit 14c estimates the genre and the domain by using the time point when the current domain input by the user is received as the purpose clarity estimation time. The clarity acquisition unit 14d acquires the objective clarity.

  The information selection unit 15 selects information to be provided to the user terminal based on the purpose clarity acquired by the goal clarity acquisition unit 14d. Specifically, when the obtained purpose clarity is equal to or greater than a predetermined threshold, the information selection unit 15 stores, for example, a web page including a feature word extracted by the estimation unit 14c to estimate the current genre. The acquisition unit 13 is searched and the URL of the searched web page is transmitted to the corresponding user terminal.

  In addition, when the acquired object clarity is less than a predetermined threshold, the information selection unit 15 includes, for example, a related word related to the feature word together with the feature word extracted to estimate the current genre. The content acquisition unit 13 is searched for the web page, and the URL of the searched web page is transmitted to the corresponding user terminal.

  Alternatively, the information selection unit 15 changes the mixing ratio between the URL searched using the characteristic word as a keyword and the URL searched using the related word as a keyword in accordance with the acquired purpose clarity, to the user terminal. Send.

  Subsequently, a processing procedure of the information providing apparatus 10 according to the first embodiment will be described with reference to FIGS. 8 and 9. FIG. 8 is a flowchart for explaining the correspondence table determination process of the information providing apparatus in the first embodiment, and FIG. 9 is a flowchart for explaining the information providing process of the information providing apparatus in the first embodiment.

  As illustrated in FIG. 8, when the known data including “domain, genre, purpose clarity” is input to the information providing apparatus 10 according to the first embodiment (Yes in step S101), the known data is stored in the correspondence table storage unit 14a. Store (step S102).

  Then, the data interpolation unit 14b determines whether there is unknown data in the stored correspondence table (step S103).

  Here, when there is no unknown data in the stored correspondence table (No at Step S103), the data interpolation unit 14b determines that the correspondence table is confirmed, and ends the processing.

  On the other hand, when there is unknown data in the stored correspondence table (Yes in step S103), the data interpolation unit 14b interpolates the unknown data using the correspondence table of the same user or the correspondence table of similar users. The table is fixed (step S104), and the process is terminated.

  When the determination process for the correspondence table is completed, the information providing apparatus 10 according to the first exemplary embodiment starts the information providing process.

  That is, as illustrated in FIG. 9, when the information providing apparatus 10 according to the first embodiment becomes the above-described purpose clarity estimation time (Yes in Step S201), the estimation unit 14c determines the current domain and genre of the user from the action log. Is estimated (step S202).

  Then, the purpose clarity level acquisition unit 14d searches the correspondence table using the domain and genre estimated by the estimation unit 14c and acquires the goal clarity level (step S203).

  After that, the information selection unit 15 selects information based on the purpose clarity acquired by the goal clarity acquisition unit 14d, provides the information to the corresponding user terminal (step S204), and ends the process.

  Note that the processing described with reference to FIGS. 8 and 9 is executed for all users who operate the user terminals constituting the user terminal group 20.

  As described above, in the first embodiment, the correspondence table storage unit 14a stores known data including “domain, genre, and purpose clarity”, and the data interpolation unit 14b stores unknown data in the stored correspondence table. In some cases, the correspondence table of the same user or the correspondence table of similar users is used to interpolate unknown data to determine the correspondence table. Then, the estimation unit 14c estimates the current domain and genre of the user from the action log, and the purpose clarity level acquisition unit 14d searches the correspondence table using the domain and genre estimated by the estimation unit 14c and clearly identifies the purpose. Get the degree. And the information selection part 15 selects information based on the objective clarity acquired by the objective clarity acquisition part 14d, and provides it to a corresponding user terminal.

  Therefore, the information selected according to the purpose clarity is information that is narrowed down within the range desired by the user without being within the range expected by the user or similar information. Therefore, it is possible to provide information that arouses new interest, and as described above, it is possible to prevent the user's satisfaction with the provided information from decreasing.

  In addition, since the correspondence table is determined by interpolation processing using machine learning, the user does not necessarily have to complete the correspondence table, and the burden on the user when estimating the purpose clarity can be reduced.

  In addition to providing information, it is possible to investigate the user's past behavior and attendance status, and to estimate the behavior when the user is acting, by estimating the purpose clarity for each user. It becomes possible to detect a user who violates the security policy.

  In the above-described first embodiment, the case where the target clarity is estimated using the correspondence table determined by machine learning has been described. In the second embodiment, the target clarity is determined using a conversion function determined by machine learning. The case of estimation will be described with reference to FIG. FIG. 10 is a block diagram illustrating a configuration of the purpose clarity degree estimation unit according to the second embodiment.

  The information providing apparatus 10 according to the second embodiment has the same configuration as the information providing apparatus 10 according to the first embodiment described with reference to FIG. 2, but the configuration of the purpose clarity degree estimation unit 14 is different from that of the first embodiment.

  In other words, as shown in FIG. 10, the purpose-clarity estimation unit 14 in the second embodiment includes a conversion function storage unit 14e, an action execution time calculation unit 14f, and a purpose-clarity acquisition unit 14d.

  The conversion function storage unit 14e stores a conversion function determined by machine learning in order to convert action execution time into purpose clarity.

  For example, the administrator of the information management device 10 performs a questionnaire survey on the user and collects the purpose clarity in each action log. Then, the administrator performs regression analysis using the behavioral execution time as an explanatory variable, using the behavioral execution time obtained by subtracting the start time from the end time as the teacher data. Determine the transformation function with 従 属 as the dependent variable. The conversion function is determined for each user. Further, the conversion function determination process by regression analysis may be executed by the information providing apparatus 10 or may be executed by a computer other than the information providing apparatus 10.

  The action execution time calculation unit 14f calculates “action execution time” from “start time” and “end time” (see FIG. 4) in the action log.

  The purpose clarity level acquisition unit 14d according to the second embodiment acquires the goal level of clarity by substituting the “behavior execution time” calculated by the behavior execution time calculation unit 14f into the conversion function stored in the conversion function storage unit 14e. .

  Note that the behavior execution time calculation unit 14f calculates the behavior execution time, and the purpose clarity level acquisition unit 14d acquires the goal clarity level (purpose clarity level estimation timing). , That is, a time when a new action log is stored and a time when the current position of the user changes.

  As the purpose clarity estimation timing, if the estimation unit 14c that executes the browsing content clustering process described in the first embodiment is also provided in the second embodiment, the third timing described above, that is, the user's request changes. It may be the case when it is time.

  Also, the “behavior execution time” assigned to the conversion function may be “behavior execution time” calculated from the action log of the purpose clarity degree estimation timing, It may be “average action execution time” calculated from the action log stored most recently in the action log. For example, “behavior execution time” is calculated from each action log of the purpose clarity estimation time from the action log stored 5 minutes before the purpose clarity estimation time, and the average value of the calculated “action execution time” is converted into a conversion function The purpose clarity may be calculated by substituting for.

  The information selection unit 15 selects information to be provided to the user terminal based on the purpose clarity acquired by the goal clarity acquisition unit 14d. For example, the information selection unit 15 in the second embodiment has the feature word extraction function of the estimation unit 14c described in the first embodiment, and when the obtained purpose clarity is equal to or higher than a predetermined threshold, the current behavior A feature word is extracted from the web page text corresponding to the log, the content acquisition unit 13 is searched for a web page including the extracted feature word, and the URL of the searched web page is transmitted to the corresponding user terminal.

  In addition, when the acquired object clarity is less than a predetermined threshold, the information selection unit 15 causes the content acquisition unit 13 to search for a web page including a related word related to the feature word together with the extracted feature word, The URL of the searched web page is transmitted to the corresponding user terminal.

  Alternatively, the information selection unit 15 changes the mixing ratio between the URL searched using the characteristic word as a keyword and the URL searched using the related word as a keyword in accordance with the acquired purpose clarity, to the user terminal. Send.

  Next, a processing procedure of the information providing apparatus 10 according to the second embodiment will be described with reference to FIG. FIG. 11 is a flowchart for explaining information providing processing of the information providing apparatus according to the second embodiment.

  As illustrated in FIG. 11, when the information providing apparatus 10 according to the second embodiment becomes the above-described purpose clarity estimation timing (Yes in step S301), the action execution time calculating unit 14f sets the “start time” and “end” in the action log. The “action execution time” is calculated from the “time” (step S302).

  Then, the purpose clarity level acquisition unit 14d substitutes the action execution time calculated by the behavior execution time calculation unit 14f into the conversion function stored in the conversion function storage unit 14e, and acquires the purpose clarity level (step S303).

  After that, the information selection unit 15 selects information based on the purpose clarity acquired by the goal clarity acquisition unit 14d, provides the information to the corresponding user terminal (step S304), and ends the process.

  Note that the processing described with reference to FIG. 11 is executed for all users who operate the user terminals constituting the user terminal group 20.

  As described above, in the second embodiment, as in the first embodiment, the information selected according to the purpose clarity is information narrowed down within the range desired by the user. As a result, the user has a new interest. Can be provided, and it is possible to prevent the user's satisfaction with the provided information from being lowered.

  In addition, because the purpose clarity is estimated using a transformation function determined by regression analysis using machine learning, etc., it is not necessary for the user to enter the goal clarity in all action logs, and the goal clarity is estimated. It is possible to reduce the burden on the user.

  In the second embodiment described above, the case where the purpose clarity is estimated using the conversion function having the action execution time as the explanatory variable has been described. However, in the third embodiment, the conversion using the value different from the action execution time as the explanatory variable. A case where the purpose clarity is estimated using a function will be described with reference to FIG. FIG. 12 is a block diagram illustrating a configuration of a purpose clarity degree estimation unit according to the third embodiment.

  The information providing apparatus 10 in the third embodiment has the same configuration as that of the information providing apparatus 10 in the first embodiment described with reference to FIG. 2, but the configuration of the purpose clarity estimation unit 14 is the same as that in the first and second embodiments. Is different.

  That is, as shown in FIG. 12, the purpose-clarity estimation unit 14 according to the third embodiment includes a conversion function storage unit 14e, a genre change degree calculation unit 14g, and a purpose-clarity acquisition unit 14d.

  First, the genre change degree used in the present embodiment will be described. As described in the first embodiment, the “genre” is a keyword representing a field to which the user's action content belongs, and is estimated from a feature word extracted from a web page text viewed by the user. The “genre change degree” is the degree to which the genre changes in the action log along a series of time series of the same user.

  Therefore, the genre change degree calculation unit 14g estimates the user's genre from the action log in the same manner as the estimation unit 14c in the first embodiment, and calculates the genre change degree by one of the three calculation methods described below. .

  As a first calculation method, the genre change degree calculation unit 14g calculates a genre similarity between adjacent action logs in a series of action logs, and sets the calculated similarity as a genre change degree.

  As a second calculation method, the genre change degree calculation unit 14g analyzes the web page text corresponding to each “URL” of the series of action logs as described in the first embodiment, and the similarity between the web page texts is analyzed. Based on the degree, the browsing contents of the user are clustered for each “request”. Then, the genre change degree calculation unit 14g calculates the number of action logs included in one request as the genre change degree.

  As a third calculation method, the genre change degree calculation unit 14g clusters user browsing contents for each “request” as in the second calculation method. And the genre change degree calculation part 14g calculates the similarity degree of the genre between the action logs contained in one request | requirement, and makes the calculated similarity degree a genre change degree.

  The conversion function storage unit 14e according to the third embodiment stores a conversion function determined by machine learning in order to convert a genre change degree into a target clarity.

  For example, the administrator of the information management device 10 performs a questionnaire survey on the user and collects the purpose clarity in each action log. Then, the administrator explains the genre change degree by performing regression analysis using the genre change degree calculated in advance by the genre change degree calculating unit 14g and the purpose clarity in each action log input by the user as teacher data. A conversion function is determined that has a variable and a target clarity as a dependent variable. The conversion function is determined for each user. Further, the conversion function determination process by regression analysis may be executed by the information providing apparatus 10 or may be executed by a computer other than the information providing apparatus 10.

  The purpose clarity obtaining unit 14d in the third embodiment substitutes the current “genre change degree” calculated by the genre change degree calculation unit 14g into the conversion function stored in the conversion function storage unit 14e in the third embodiment. , Get purpose clarity.

  The genre change degree calculation unit 14g calculates the genre change degree, and the purpose clarity degree acquisition part 14d acquires the purpose clarity degree (purpose clarity degree estimation timing) as described above. For example, the time when a new action log is stored, the time when the current position of the user changes, and the like.

  The information selection unit 15 selects information to be provided to the user terminal based on the purpose clarity acquired by the goal clarity acquisition unit 14d. Specifically, the information selection unit 15 according to the third exemplary embodiment extracts the feature word extracted by the genre change degree calculation unit 14g to estimate the current genre when the obtained purpose clarity is equal to or greater than a predetermined threshold, for example. The content acquisition unit 13 searches for the web page including the URL of the searched web page and transmits it to the corresponding user terminal.

  In addition, the information selection unit 15 according to the third embodiment, for example, when the acquired purpose clarity is less than a predetermined threshold, for example, together with the feature word extracted by the genre change degree calculation unit 14g to estimate the current genre, The content acquisition unit 13 is made to search for a web page including a related word related to the feature word, and the URL of the searched web page is transmitted to the corresponding user terminal.

  Alternatively, the information selection unit 15 changes the mixing ratio between the URL searched using the characteristic word as a keyword and the URL searched using the related word as a keyword in accordance with the acquired purpose clarity, to the user terminal. Send.

  Subsequently, a processing procedure of the information providing apparatus 10 according to the third embodiment will be described with reference to FIG. FIG. 13 is a flowchart for explaining information providing processing of the information providing apparatus according to the third embodiment.

  As illustrated in FIG. 13, when the information providing apparatus 10 according to the third embodiment becomes the above-described purpose clarity estimation time (Yes in step S401), the genre change degree calculation unit 14g determines the user's current genre from the action log. After estimation, the current genre change is calculated (step S402).

  Then, the purpose clarity degree acquisition unit 14d substitutes the genre change degree calculated by the genre change degree calculation part 14g into the conversion function stored in the conversion function storage part 14e, and acquires the purpose clarity degree (step S403).

  After that, the information selection unit 15 selects information based on the purpose clarity acquired by the goal clarity acquisition unit 14d, provides the information to the corresponding user terminal (step S404), and ends the process.

  Note that the processing described with reference to FIG. 13 is executed for all users who operate the user terminals constituting the user terminal group 20.

  As described above, in the third embodiment, as in the first and second embodiments, the information selected in accordance with the purpose clarity is information narrowed down within the range desired by the user. It is possible to provide information that arouses interest and to prevent the user's satisfaction with the provided information from decreasing.

  In addition, the purpose clarity is estimated using the genre change automatically calculated by the genre change calculation unit 14g and a conversion function determined by regression analysis using machine learning, etc. It is not necessary to input the purpose clarity in the log, and the burden on the user when estimating the goal clarity can be reduced.

  In the above-described first to third embodiments, the case where the information providing apparatus 10 is interposed between the user terminal group 20 and the content server group 30 and provides information based on the purpose clarity is described. However, the information providing apparatus 10 may be set for each user terminal constituting the user terminal group 20 or for each content server constituting the content server group 30.

  That is, when installed in a user terminal operated by a plurality of users, the information providing apparatus 10 obtains purpose clarity for each user operating the user terminal and selects and provides information. Further, when installed in the content server, the information providing apparatus 10 obtains purpose clarity for each user operating the user terminal that has made a distribution request to the installed content server, and selects and provides information. To do.

  In addition, among the processes described in this embodiment, all or part of the processes described as being performed automatically can be performed manually, or the processes described as being performed manually can be performed. All or a part can be automatically performed by a known method. In addition, the processing procedure, control procedure, specific name, and information including various data and parameters shown in the above-described document and drawings can be arbitrarily changed unless otherwise specified.

  Each component of each illustrated device is functionally conceptual and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution / integration of each device is not limited to that shown in the figure, and all or a part thereof may be functionally or physically distributed or arbitrarily distributed in arbitrary units according to various loads or usage conditions. Can be integrated and configured. Further, all or any part of each processing function performed in each device may be realized by a CPU and a program analyzed and executed by the CPU, or may be realized as hardware by wired logic.

  The information selection index estimation method and the information provision method described in this embodiment can be realized by executing a program prepared in advance on a computer such as a personal computer or a workstation. This program can be distributed via a network such as the Internet. The program can also be executed by being recorded on a computer-readable recording medium such as a hard disk, a flexible disk (FD), a CD-ROM, an MO, and a DVD and being read from the recording medium by the computer.

  As described above, the information selection index estimation method, the information selection index estimation device, the information selection index estimation program, the information provision method and the information provision system according to the present invention are useful when selecting information desired by the user, In particular, it is suitable for preventing the user's satisfaction with the provided information from decreasing.

DESCRIPTION OF SYMBOLS 10 Information provision apparatus 11 Action log acquisition part 12 Action log memory | storage part 13 Content acquisition part 14 Purpose clarity degree estimation part 14a Correspondence table memory | storage part 14b Data interpolation part 14c Estimation part 14d Purpose clarity degree acquisition part 14e Conversion function memory | storage part 14f Action execution Time calculation unit 14g Genre change degree calculation unit 15 Information selection unit 20 User terminal group 30 Content server group

Claims (9)

An information selection index estimation method for estimating an information selection index used to select information desired by a user,
An action history storage step of storing the user's action history in an action history storage unit;
Based on the behavior history stored in the behavior history storage unit, the purpose clarity is a degree indicating a dependency relationship between the information intended by the user at the present time and the topic that the user is currently looking at. An acquisition step to acquire as an information selection index;
The information selection parameter | index estimation method characterized by including. A correspondence table storage step of storing a correspondence table in which a position at the time of action execution, a field of action content, and the purpose clarity are associated with each user in the correspondence table storage unit;
From the behavior history stored in the behavior history storage unit, an estimation step for estimating the position of the user who performed the behavior corresponding to the behavior history at the time of the behavior execution and the field of the behavior content corresponding to the behavior history;
Further including
The acquisition step searches the correspondence table of the user stored in the correspondence table storage unit for the purpose clarity corresponding to the field of the user's behavior execution and the field of action content estimated by the estimation step. The information selection index estimation method according to claim 1, wherein the information selection index estimation method is obtained.   In the correspondence table stored in the correspondence table storage unit, when the purpose clarity corresponding to the position at the time of the action execution and the field of the action content is uncertain, the uncertain purpose clarity is interpolated by machine learning. The information selection index estimation method according to claim 2, further comprising an interpolation step. A conversion function storing step of storing in the conversion function storage unit a conversion function determined by machine learning to convert the action execution time into the objective clarity.
The action history storage step further stores an action start time and an action end time in association with the action history in the action history storage unit,
In the obtaining step, the action execution time calculated from the action start time and action end time associated with the action history stored in the action history storage unit is substituted into the conversion function stored in the conversion function storage unit. The information selection index estimation method according to claim 1, wherein the purpose clarity is acquired. A conversion function storage step of storing, in a conversion function storage unit, a conversion function determined by machine learning in order to convert a field change degree that is a change degree along a time series of the field of action content into the purpose clarity;
A field change degree calculating step of estimating a field change degree by estimating a field of the action content from the action history stored in the action history storage unit;
The acquisition step further includes: The information selection index estimation method according to claim 1, wherein the information selection index is estimated. An information selection index estimation device that estimates an information selection index used to select information desired by a user,
Action history storage means for storing the user's action history;
Based on the behavior history stored in the behavior history storage means, the purpose clarity is a degree indicating a dependency relationship between the information intended by the user at the present time and the topic currently being watched by the user. Acquisition means for acquiring as an information selection index;
An information selection index estimation apparatus comprising: An information selection index estimation program for causing a computer to execute an information selection index estimation method for estimating an information selection index used for selecting information desired by a user,
An action history storage procedure for storing the user's action history in an action history storage unit;
Based on the behavior history stored in the behavior history storage unit, the purpose clarity is a degree indicating a dependency relationship between the information intended by the user at the present time and the topic that the user is currently looking at. An acquisition procedure to acquire as an information selection index;
An information selection index estimation program for causing a computer to execute. An information providing method applied to the information providing apparatus in an information providing system including a user terminal operated by a user and an information providing apparatus that selects and provides information desired by the user to the user terminal,
An action history storage step of storing the user's action history in an action history storage unit;
Based on the behavior history stored in the behavior history storage unit, a purpose clarity level that is a degree indicating a dependency relationship between information that the user is currently interested in and a topic that the user is currently watching is acquired. An acquisition step to
An information selection step of selecting information to be provided to the user terminal based on the purpose clarity acquired by the acquisition step;
An information providing method characterized by including An information providing system composed of a user terminal operated by a user and an information providing apparatus that selects and provides information desired by the user to the user terminal,
The information providing apparatus includes:
Action history storage means for storing the user's action history;
Based on the behavior history stored in the behavior history storage means, a purpose clarity level that is a degree indicating a dependency relationship between information that the user aims at the present time and a topic that the user is currently looking at is acquired. Acquisition means to
Information selecting means for selecting information to be provided to the user terminal based on the purpose clarity obtained by the obtaining means;
An information providing system characterized by comprising:

Images