JP2009129057A - Information recommendation system, information recommendation method, and program for information recommendation - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information recommendation system, an information recommendation method and an information recommendation program for recommending information leading to the action of a user in a proper timing. <P>SOLUTION: This information recommendation system is provided with a content determination means for determining content to be recommended as content for presentation based on the combination of a browsing date when a user has browsed content and an action date when the user has acted based on the browsed content, and for determining a period when the content is recommended. Thus, it is possible to recommend information leading to the action of the user in a proper timing. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an information recommendation system, an information recommendation method, and an information recommendation program for recommending information such as content to a user.

  Patent Document 1 describes an example of an advertisement system that presents an effective advertisement based on a time when a user receives information (reception time) and a position (reception position) where the information is received. The advertisement system described in Patent Literature 1 includes an advertisement database, a time acquisition unit, a preference acquisition unit, a position acquisition unit, an advertisement storage processing unit, a received advertisement database, an advertisement extraction unit, and an advertisement presentation unit. With.

  The advertising system described in Patent Literature 1 operates as follows, for example. That is, the advertisement database stores advertisement data to which an advertisement presentation period, an advertisement validity period, an advertisement target business time, and an advertisement target position are assigned. The time acquisition unit acquires the current time. The preference acquisition unit acquires the user's preference. The position acquisition unit acquires the current position of the user.

  The advertisement storage processing unit sets selection criteria for the reception time, the reception position of the user, and the preference. For example, the selection criteria are that the reception time is within the advertisement period, that the reception position is within a certain distance from the advertisement target position, and that no keyword that the user is not interested in is included in the advertisement data. Then, the advertisement storage processing unit selects advertisement data satisfying the selection criteria from the received advertisement data and stores it in the received advertisement database.

  The advertisement extraction unit evaluates the advertisement data stored in the received advertisement database based on the playback time, the playback position of the user, and the preference, and extracts them in descending order of evaluation. As a method of evaluating advertisement data, for example, it is higher when the reception time is within the advertisement period, and the shorter the distance between the reception position and the position of the advertisement target, the higher, and the advertisement data includes many keywords that are of interest to the user. There is a way to appreciate it so much. The advertisement presenting unit presents the extracted advertisement data to the user.

  Further, in Patent Document 2, as an element for evaluating the tendency of the user's interest, in addition to the past document browsing history, at least one piece of information on the voice information including the user's behavior history and the user's conversation is used. Thus, there is described an information presentation system that can determine a tendency of interest of the user and efficiently present appropriate information.

JP 2004-240986 A JP 2004-272355 A

  The problem of the technique described in Patent Document 1 is that information that matches or closes the time and position at which the user browses the information with the period of the advertisement and the position of the advertisement is presented. It is. The reason for this is that, generally, the reception time is within the advertisement period, and the short distance between the reception position and the advertisement target position increases the points when selecting or evaluating advertisement data, and within the advertisement period. This is because advertisement data that is closer in position is easier to extract.

  However, the user may now seek information about a place near here at this location, but often asks for information in advance for some future action. The technique described in Patent Document 1 cannot cope with a case where the timing for acquiring information differs from the timing for using information.

  Further, in the method described in Patent Document 2, content related to a place where the user has acted in the past is determined as content having a high degree of interest of the user by determining the relevance between the action history and the presentation schedule information (content). Present. However, the time of action, that is, the timing of taking action is not considered. Further, the timing of presenting content is not taken into consideration. Therefore, it is not always possible to present content at an appropriate timing.

  Therefore, an object of the present invention is to provide an information recommendation system, an information recommendation method, and an information recommendation program that can recommend information related to user behavior at an appropriate timing.

  The information recommendation system according to the present invention determines the content to be recommended as the presentation content based on the combination of the browsing date and time when the content is browsed and the action date and time when the user acts based on the browsed content, and the time when the content is recommended The content determining means for determining the content is provided.

  The information recommendation method according to the present invention determines the content to be recommended as the presentation content based on the combination of the browsing date and time when the content is browsed and the action date and time when the user acts based on the browsed content. A content determination step for determining the content.

  The information recommendation program according to the present invention determines the content to be recommended as the presentation content based on the combination of the browsing date and time when the content is browsed and the action date and time when the computer acts based on the browsed content. It is characterized in that content determination processing for determining when to recommend is executed.

  According to the present invention, there is an effect that information that leads to user behavior can be recommended at an appropriate timing.

  First, an outline of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a minimum configuration example of an information recommendation system according to the present invention. The information recommendation system 10 illustrated in FIG. 1 includes content determination means 404.

  The content determination unit 404 determines the time to recommend the content and the recommended content based on the combination of the browsing date and time when the content was browsed and the action date and time when the user acted based on the browsed content.

  If it comprises as shown in FIG. 1, the information which leads to a user's action can be recommended at an appropriate timing.

Embodiment 1. FIG.
Next, a first embodiment of the present invention will be described with reference to the drawings. FIG. 2 is a block diagram illustrating a configuration example of the information recommendation system according to the first embodiment. Referring to FIG. 2, the first embodiment of the present invention includes an input means 1, a display means 2, a storage means 3 for storing information, and a data processing means 4 that operates by program control.

  The information recommendation system according to the present invention obtains the relevance between the action history and the content that has already been browsed based on whether or not the user actually went to the location indicated by the content that the user browsed. In addition, the analysis is performed by paying attention to the correlation between the browsing date and the action date and time, the user will predict when to act next, and the content suitable for the future action is determined as the recommended content. “Determine as recommended content” means to determine content to be presented to the user by display or the like.

  The input unit 1 is realized by, for example, a CPU, a key operation unit, a GPS receiver, and the like included in an information terminal such as a mobile phone that is carried by a user and operates according to a program. When the information recommendation system is realized by a single device, it is realized by a CPU, a keyboard, etc. of an information processing device such as a personal computer. The input unit 1 includes a content selection unit 101 and an action acquisition unit 102.

  The display unit 2 is realized by, for example, a CPU and a display display unit of an information terminal or an information processing device. The display unit 2 includes a content list display unit 201 and a content content display unit 202.

  The storage means 3 is realized by, for example, a database device such as a magnetic disk device or a database server. The storage unit 3 includes a content database (DB) 301, a browsing history database (DB) 302, an action history database (DB) 303, and a browsing action comparison database (DB) 304.

  The data processing means 4 is realized by an information processing apparatus such as a personal computer that operates according to a program, for example. The data processing unit 4 includes a browsing behavior comparison unit 401, a ratio calculation unit 402, a priority determination unit 403, and a content determination unit 404.

  The content database 301 stores one or more contents, and information indicating a location, a period, and a genre regarding the content target is assigned to each content.

  The content list display unit 201 displays a plurality of contents obtained as a result of searching the content database 301 under a certain condition. For example, the content list display unit 201 displays a screen for allowing the user to select one content from a plurality of contents searched from the content database 301.

  The content selection unit 101 receives an instruction for the user to select one content from the content displayed by the content list display unit 201. That is, the content selection unit 101 outputs, for example, information that can identify the content selected by the user to the content content display unit 202 together with information that can identify the user (hereinafter referred to as a user ID).

  The content content display unit 202 refers to the content database 301 and displays the content of one content selected by the user, and transmits (or outputs) necessary information to the browsing history database 302.

  Specifically, the content content display unit 202 extracts content from the content database 301 based on information that can identify the content selected by the user, displays the extracted content, and allows the user to browse. In addition, the content content display unit 202 stores, as necessary information, the user ID, information indicating the date and time when the user browsed the content (browsing date and time), and information indicating the location given to the content in the browsing history database 302.

  The browsing history database 302 stores, for each user ID, information indicating the date and time (browsing date and time) when the user browsed the content and the location assigned to the content. That is, the browsing history database 302 stores information transmitted by the content content display unit 202.

  The action acquisition unit 102 acquires information indicating the user's action and transmits (or outputs) necessary information to the action history database 303. In addition, the action acquisition unit 102 obtains user position information based on the GPS signal as information indicating the action of the user, for example. The action acquisition unit 102 stores, as necessary information, a user ID, information indicating the date and time when the user actually acted (action date and time), and information indicating the place where the action was performed in the action history database 303.

  The behavior history database 303 stores the date and time when the user actually acted (behavior date and time) and information indicating the place where the user went for each user ID.

  The browsing behavior comparison unit 401 analyzes the correlation between the browsing history database 302 and the behavior history database 303 and creates a browsing behavior comparison database 304.

  For example, the browsing behavior comparison unit 401 creates the browsing behavior comparison database 304 by the following processing. For example, the browsing behavior comparison unit 401 compares information indicating a location stored in the browsing history database 302 with information indicating a location performed by the user stored in the behavior history database 303. Of the information stored in the browsing history database 302 and the behavior history database 303, information to be compared is information corresponding to the same user ID, for example.

  When the compared information matches, the browsing behavior comparison unit 401 extracts the browsing date and time corresponding to the information indicating the matching location from the browsing history database 302. Similarly, the action date and time corresponding to the information indicating the matched location is extracted from the action history database 303.

  The browsing behavior comparison unit 401 stores the extracted browsing date / time and behavior date / time in the browsing behavior comparison database 304 in association with the user ID and the information indicating the matched location.

  The ratio calculation means 402 calculates the ratio that the combination appears in the entire combination stored in the browsing action comparison database 304 with the browsing date and the action date as a set. Further, the ratio calculation unit 402 stores the calculated result in the browsing behavior comparison database 304.

  The browsing behavior comparison database 304 stores, for each piece of information indicating the user ID and the matched location, a set of browsing date and time and behavior date and a ratio at which the combination appears with respect to the entire combination stored in the database.

  The priority determination means 403 determines the priority for recommending “information for which” for each browsing date and time based on the ratio calculated by the ratio calculation means 402 and stored in the browsing behavior comparison database 304.

  The priority is a value for determining recommended content and is a value for the action date and time. For example, the priority determination unit 403 has a higher priority on the action date and time included in the information having the higher “ratio” in which the combination of the view date and time and the action date and time appear in the information with the same view date and time stored in the browse action comparison database 304. Assign.

  In addition, the priority determination unit 403 determines a time based on the browsing date and the action date as the time to recommend the content in order to recommend the content at a timing easily associated with the action. For example, the priority determination unit 403 sets the browsing date / time in the combination of the browsing date / time and action date / time having a high priority as the time to recommend the content.

  The content determination unit 404 determines the recommended content based on the priority determined by the priority determination unit 403. For example, the content determination unit 404 determines the content related to the action date with high priority as the recommended content. For example, the content determination unit 404 determines, from the content database 301, content that includes a high-priority action date and time in the “period” as recommended content.

  Next, the operation of the first embodiment will be described with reference to the drawings. FIG. 3 is a flowchart showing an example of the operation of the information recommendation system in the first embodiment.

  First, the content list display unit 201 displays a plurality of contents obtained as a result of searching the content database 301 under a certain condition (step S101). When the user selects one content (step S102), the content selection unit 101 outputs, for example, a user ID and information that can identify the content selected by the user to the content content display unit 202. The content content display means 202 refers to the content database 301 and displays the content of one content selected by the user (step S103) and transmits necessary information to the browsing history database 302 (step S104).

  Here, the operation of the action acquisition unit 102 will be described. FIG. 4 is a flowchart illustrating an example of the operation of the action acquisition unit 102. For example, the action acquisition unit 102 executes the process illustrated in FIG. 4 as needed.

  When the user takes an action, the action acquisition unit 102 acquires information indicating the action (step S201) and transmits necessary information to the action history database 303 (step S202).

  Referring to FIG. 3, the browsing behavior comparison unit 401 reads the browsing history database 302 and the behavior history database 303 (step S105), converts the location information expressed in a place other than the latitude and longitude into latitude and longitude (step S106), The correlation between the two is analyzed (step S107), and the browsing behavior comparison database 304 is created (step S108).

  That is, in step S106, the browsing behavior comparison unit 401 stores the “information indicating the location given to the content” stored in the browsing history database 302 and the “information indicating the location performed by the user” stored in the behavior history database 303. The information is converted into the same format that can be correlated.

  The ratio calculation means 402 calculates the ratio at which the combination of the browsing date / time and the behavior date / time appears in the entire browsing behavior comparison database 304 (step S109). The priority determination unit 403 determines the priority for recommending “information for which” for each browsing date based on the ratio calculated by the ratio calculation unit 402 (step S110). Finally, the content determination unit 404 determines the recommended content based on the priority determined by the priority determination unit 403 (step S111).

  Next, the effect of the first embodiment will be described. In the first embodiment, since the content is recommended to the user at a timing (timing easily linked to the user's behavior) with a strong relationship between the browsing timing and the behavior timing, the recommended content is the behavior. The probability that it will be used is increased. That is, it is possible to reduce the probability that the recommended content is wasted without being used by the user.

  The recommended content is not determined based on the user's current location, but is analyzed to determine when there is a tendency to act based on the content being browsed. When they want to obtain information in advance for future actions, they can provide useful information at the appropriate time.

  Hereinafter, another preferable aspect of the information recommendation system according to the present invention will be described. The information recommendation system of the present invention stores one or more contents, and each content is given a location, period, and genre for the content target, and the content database is referred to and the user is referred to. Displays the content of one content selected by the user, and stores the content content display means for transmitting necessary information to the browsing history database, the date and time when the user browsed the content, and the location assigned to the content for each user. Browsing history database, behavior acquisition means for acquiring user behavior and transmitting necessary information to the behavior history database, behavior history for storing the date and time when the user actually acted, and the place where the user has performed for each user Database, browsing history database and action history database A browsing behavior comparison means for analyzing the relation and creating a browsing behavior comparison database, a ratio calculation means for calculating a ratio of the browsing date and action date as a set, and the combination appearing in the entire browsing behavior comparison database, and a user For each place, a set of browsing date and action date, and a browsing behavior comparison database that memorizes the ratio of the combination appearing in the entire database, and recommending "information for when" for each browsing date based on the ratio Priority determining means for determining whether or not to perform content, and content determining means for determining recommended content based on the priority.

  By adopting such a configuration, the degree of connection between the timing when the user browses the content and the timing when the user has acted based on the content is obtained, and the content is recommended by giving priority to the timing that is strongly related. Can achieve the purpose.

Embodiment 2. FIG.
Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 5 is a block diagram illustrating a configuration example of the information recommendation system according to the second embodiment. Referring to FIG. 5, in addition to the first embodiment, the second embodiment of the present invention includes data acquisition means 5 that operates under program control.

  The data acquisition unit 5 includes an external data acquisition unit 501, an external data conversion unit 502, and an external action history database (DB) 503.

  The external data acquisition unit 501 acquires an action history from other than the input unit 1 and registers it in the external action history database 503. As an action history acquired from other than the input means 1, for example, a purchase history from a credit card company or a transit record of transportation is conceivable. There are various data acquisition methods depending on the target, such as via a network or reading from a medium such as a CD. FIG. 6 is an explanatory diagram showing a registration example of the external action history database 503. In the example shown in FIG. 6, a purchase history with a credit card is registered as an action history.

  The external data conversion unit 502 converts the data registered in the external action history database 503 and updates the contents of the external action history database 503. In general, the action history acquired from the outside has a format different from that of data registered in the action history database 303. Therefore, the external data conversion means 502 converts the action history acquired from the outside into a predetermined format. For example, it is possible to associate user IDs, add location information based on the store name, and associate purchase types with genres. Various databases necessary for conversion are also included in the external data conversion means 502.

  In the second embodiment, the browsing behavior comparison unit 401 reads the external behavior history database 503 together with the browsing history database 302 and the behavior history database 303. The browsing behavior comparison unit 401 uses the external behavior history database 503 in the same manner as the behavior history database 303.

  According to the second embodiment, the browsing behavior comparison database 304 can be created based on more behavior histories. Therefore, it is possible to determine the content to be recommended more accurately.

  Next, a first embodiment of the present invention will be described with reference to the drawings. Such an example corresponds to the first embodiment. Hereinafter, the case where the information recommendation system according to the first embodiment has the configuration shown in FIG. 2 will be described as an example.

  In the first embodiment, the information recommendation system includes, for example, a CPU, a keyboard, a GPS (Global Positioning System, etc.) included in an information terminal such as a mobile phone carried by a user or an information processing device such as a personal computer as an input means 1. Global Positioning System) equipped with a receiver. The data processing unit 4 includes a personal computer, and the storage unit 3 includes a magnetic disk device. Further, the display means 2 includes, for example, an information terminal, a CPU of an information processing device, and a display device. The personal computer has a central processing unit. Various databases are stored on the magnetic disk.

  The content database 301 stores one or more contents, and information indicating a location, a period, and a genre regarding the content target is assigned to each content. FIG. 7 is an explanatory diagram showing a registration example of the content database 301. FIG. 7 illustrates a case where the content database 301 stores information on [content URL], [location], [period], [genre], and [text]. [Location] is, for example, an address, area, nearest station, latitude / longitude, and store name. [Period] is, for example, business hours or event periods such as sales. [Genre] is, for example, the type of content such as gourmet food, sports, medical care, and English.

  The content list display unit 201 displays a list of a plurality of contents obtained as a result of searching the content database 301 with the keyword “restaurant”, for example. The content to be displayed may be a URL of the searched content, a title, or a summary. Also, a combination thereof may be used. For example, the content list display unit 201 displays a screen for allowing the user to select one content from a plurality of contents searched from the content database 301.

  The user decides one content to view detailed contents from among a plurality of contents displayed in a list, and clicks a selection button on the keyboard. For example, the content selection unit 101 outputs information that can identify the content selected by the user on the screen for selecting the content displayed by the content list display unit 201 to the content content display unit 202 together with the user ID. The content selection unit 101 may be a microphone or a touch panel touched with a finger or a pen.

  When the user selects one content, the content content display unit 202 refers to the content database 301 to display the content of the content selected by the user and transmits necessary information to the browsing history database 302. The necessary information is, for example, information indicating the date, day of the week, and time when the user browsed the content, the URL of the browsed content, and the location and genre of the content target. The content content display means 202 acquires the browsing date, day of the week, time, and content URL when the content selected by the user is displayed. Also, the location and genre of the content target are acquired from the content database 301.

  The browsing history database 302 stores such information for each user. FIG. 8 is an explanatory diagram showing a registration example of the browsing history database 302. In FIG. 8, the browsing history database 302 stores information on [user ID], [browsing date], [browsing day of the week], [browsing time], [content URL], [location], and [genre]. Is illustrated. In FIG. 8, the user “00000A” browses the content of the URL “www.aaa.bbb.com” at “17:00” on “Wednesday” on “05/09/2007” and The subject is “Shibaura 3-chome, Minato-ku” and the genre is “Gourmet”.

  On the other hand, behavior acquisition means 102 realized by a GPS receiver or the like frequently acquires user behavior. If you don't get it frequently, you won't know if you stayed in the same place or just passed. In addition, as the action acquisition unit 102, RFID (Radio Frequency Identification) may be used, or the user may input when the user arrives at the destination. The information acquired by the action acquisition unit 102 is, for example, information indicating a day, a day of the week, a time, a place where the user has performed, etc. The place visited is obtained by latitude and longitude in the GPS receiver, and is obtained by address or store name when the user inputs. The action acquisition unit 102 transmits the acquired information to the action history database 303.

  The action history database 303 stores such information for each user. FIG. 9 is an explanatory diagram showing an example of registration in the action history database 303. FIG. 9 illustrates a case where the action history database 303 stores information on [user ID], [action date], [action day of the week], [action time], and [location]. The example shown in FIG. 9 indicates that the user “00000A” went to “Nishino Department Store” at “17:00” on “Friday” on “2007/09/07”.

  The browsing behavior comparison unit 401 compares the browsing history database 302 and the behavior history database 303 for each user. The key for comparison is [location] stored in the browsing history database 302 and [location] stored in the behavior history database 303. If the values of both [locations] match, the browsing behavior comparison unit 401 determines that browsing and behavior are linked by a single content, that is, behavior based on the browsed content, and the browsing date / time and behavior A browsing behavior comparison database 304 in which date / location is associated is created. Note that when comparing [location] by a combination of latitude and longitude, there may be a case where the same point is not completely matched due to positioning errors and how to obtain representative points. Therefore, it is conceivable to allow a certain error when confirming a match.

  FIG. 10 is an explanatory diagram illustrating a registration example of the browsing behavior comparison database 304. In FIG. 10 (a), the browsing behavior comparison database 304 includes [user ID], [browsing day of the week], [browsing time], [place A], [acting day of the week], [acting time], and [place B]. The case where each information is memorize | stored is illustrated.

  If the [location] in the browsing history database 302 is “latitude and longitude” “111.000,222.000” and the [location] in the behavior history database 303 is also “latitude and longitude” “111.000,222.000”, the browsing behavior comparison means 401 Judging that browsing and behavior are connected. Then, the browsing behavior comparison unit 401 extracts the value “111.000, 222.000” of the “location” from the browsing history database 302 and writes it in the “location A” of the browsing behavior comparison database 304 (see FIG. 10A). . Subsequently, the browsing behavior comparison unit 401 extracts the value of “place” “111.000,222.000” from the behavior history database 303, and the value of “location A” in the browsing behavior comparison database 304 is “111.000,222.000”. Write to [Location B] on a line.

  Next, the browsing behavior comparison unit 401 extracts the values of [browsing day of the week] and [browsing time] when browsing the content whose [location] value is “111.000,222.000” from the browsing history database 302, and browsing behavior The value of [Location A] in the comparison database 304 is written in [Browse Day of Week] and [Browse Time] in the row having “111.000, 222.000”.

  Further, the browsing action comparison unit 401 extracts the values of [action day of the week] and [action time] of the action whose [place] value is “111.000, 222.000” from the action history database 303, and browse action comparison database 304. Is written in [Action Day] and [Action Time] in the row where the value of [Location B] is “111.000,222.000”.

  The values of [Place] in the browsing history database 302 and [Place] in the action history database 303 are not necessarily latitude / longitude, and may be an address, area, nearest station, or store name as described above. . Therefore, it is necessary to match the same format before comparing browsing behavior. In order to convert [location] represented by values other than latitude and longitude into latitude and longitude, a correspondence database is prepared. The following patterns and conversion methods may be considered.

  For example, when the value of [Place] is an address, for example, up to chome such as “1 chome” is extracted from the address and converted into representative latitude and longitude. If the value of [Place] is an area, the area is converted into a representative latitude / longitude. For example, when the value of [Place] is the nearest station, the nearest station is converted into latitude and longitude. Further, for example, when the value of [Place] is a store name, an address is obtained from the store name, and up to chome is extracted from the address, and converted to representative latitude and longitude.

  Even if the values of [Location] match, it may not be clear whether the action is based on browsing. For example, the value of [Place] is both an address and the address has a department store. Even if the user is viewing the contents of a restaurant in a department store, it is not clear whether the user's behavior was really eating at the restaurant or shopping at another store in the department store. is there. In such a case, the genre may be used as a key when comparing the browsing history database 302 and the action history database 303.

  Although there is an item “genre” in the browsing history database 302, it is not in the behavior history database 303. Therefore, a correspondence database for converting the value of [location] into a genre is prepared, the value of [location] in the action history database 303 is converted into a genre, and then compared with [genre] in the browsing history database 302. . The following patterns and conversion methods may be considered.

  For example, when the value of [Place] in the action history database 303 is a store name, the genre is obtained from the store name. For example, when the value of [Place] in the action history database 303 is latitude / longitude, the latitude / longitude is converted into an address, a store name is obtained from the address, and a genre is obtained from the store name.

  When the browsing behavior comparison database 304 is created and updated, the ratio calculation unit 402 sets the browsing date and the behavior date as a set, calculates a ratio at which the combination appears in the entire browsing behavior comparison database 304, and compares the browsing behavior. Write to database 304.

  In the example shown in FIG. 10A, first, the ratio calculation means 402 reads data of “Wednesday” and “Browse time” values “Wednesday” and “17:00” from the browsing behavior comparison database 304. The number (A) is calculated.

  Next, the ratio calculation means 402 has values of “browsing day of the week” and “browsing time” of “Wednesday” and “17:00”, and the values of “action day of the week” and “action time” are “Friday” and “ The number (B) of data that is “17:00” is calculated. Then, the ratio of B to A is calculated.

  Similarly, the ratio calculation means 402 has values of “browsing day of the week” and “browsing time” of “Wednesday” and “17:00”, and the values of “action day of the week” and “action time” are “Saturday” and “ The number (C) of data “13:00” is calculated, and the ratio of C to A is calculated.

  FIG. 10B is an example of the browsing behavior comparison database 304 after the ratio calculated by the ratio calculation unit 402 is written. In the example shown in FIG. 10B, the percentage that the user “00000A” goes to “17:00” on “Friday” at the place viewed at “17:00” on “Wednesday” is “20%”, “Wednesday” The ratio of going to “13:00” on “Saturday” to the place viewed at “17:00” on “Saturday” is “30%”.

  The priority determining unit 403 determines the priority of “information for which” is recommended for each browsing day and browsing time based on the ratio calculated by the ratio calculating unit 402. In the example shown in FIG. 10B, the priority determination unit 403 first sets “information for 13:00 on Saturday” and “information for 13:00 on Saturday” as the timing that is likely to lead to action. Decide to recommend "information for 17:00 on Friday" second. The recommended time for recommending the content may have a range. That is, instead of recommending exactly 17:00, for example, 16: 00-18: 00 is set to the classification of “evening”, and the information recommended for 17:00 is between 16: 00-18: 00 You may recommend it.

  Finally, the content determination unit 404 determines the recommended content based on the priority determined by the priority determination unit 403. For example, the information about the period previously given to the content is used, and the content whose scheduled action date is within the period is recommended. For example, the content determination unit 404 determines, from the content database 301, content including “Saturday 13:00”, which is an action date and time (scheduled action date) with a high priority, as recommended content. Also, using the other user's behavior history database 303, for example, “Saturday” at “13:00”, a location where the other user is frequently acting is obtained, and content related to that location is recommended. Also good.

  Next, a second embodiment of the present invention will be described with reference to the drawings. This example corresponds to the second embodiment. The case where the information recommendation system according to the second embodiment has the configuration shown in FIG. 5 will be described as an example.

  The operation of the second embodiment is the same as that of the first embodiment except for step S105 in FIG. In step S105, not only the browsing history database 302 and the behavior history database 303 but also the external behavior history database 503 are referred to. The browsing behavior comparison unit 401 uses the external behavior history database 503 in the same manner as the behavior history database 303.

  In the embodiment described above, an information recommendation system having characteristic configurations as shown in the following (1) to (5) is shown.

  (1) Content determination for determining content to be recommended as content for presentation based on a combination of browsing date and time for browsing content and action date and time for behavior based on the browsed content An information recommendation system comprising means (implemented by priority determination means 403 and content determination means 404, for example).

  (2) Among the combinations stored in the browsing behavior comparison database (for example, realized by the browsing behavior comparison database 304) that stores the combination of the browsing date and the behavior date, and the combination stored in the browsing behavior comparison database, the combination to be processed is The content determination means includes a ratio calculation means (for example, realized by the ratio calculation means 402) that calculates the proportion of the occupancy for each combination, and the content determination means sets the action date and time for each browsing date based on the ratio calculated by the ratio calculation means. An information recommendation system that extracts and determines content related to the extracted action date as recommended content. The information recommendation system configured as described above can recommend the content related to the action date / time in the combination of the browsing date / time and the action date / time that has a large proportion of the browsing action comparison database. In other words, it is possible to recommend content related to an action date and time that is strongly related to the browsing date and time.

  (3) A content database (for example, realized by the content database 301) that stores one or more contents, and stores information indicating the location, period, and genre of the content target in association with each content. , Referring to the content database, displays the content of one content selected by the user, and also displays the content details to be stored in the browsing history database (for example, realized by the browsing history database 302) indicating the browsing date and time. Means (for example, realized by the content content display means 202) and the user's action are acquired, and information indicating the action date and time is stored in the action history database (for example, realized by the action history database 303). Action acquisition means (for example, realized by action acquisition means 102) And the browsing history database and the behavior history database are analyzed, and the browsing behavior comparison unit (for example, the browsing behavior comparison unit 401 realizes a combination of the browsing date and the behavior date and time in the browsing behavior comparison database). The browsing history database stores the browsing date and time when the user browses the content and the information indicating the location given to the content for each piece of information indicating the user. The action date and time when the user actually acted and the information indicating the place where the action was performed are stored for each piece of information indicating the user. And the combination calculated by the ratio calculation means are combined with the entire combination stored in the browsing behavior comparison database. The content determination unit is realized by a priority determination unit (for example, the priority determination unit 403) that extracts the action date and time for each browsing date and determines the priority based on the ratio. ) And a content recommendation system that determines content related to the corresponding action date as recommended content based on the priority determined by the priority determination means. The information recommendation system configured as described above can determine when to recommend “information for” by analyzing the time of browsing the content and acting.

  (4) The browsing behavior comparison means compares the information indicating the location stored in the browsing history database with the information indicating the location stored in the behavior history database, and if both match, the behavior based on the browsed content The information recommendation system which judges that it is, and matches a browsing date and action date with a place as a key, and memorize | stores it in a browsing action comparison database. The information recommendation system configured as described above can combine the viewing date and time and the action date and time when the user actually goes to the location indicated by the content viewed by the user, so that the content to be recommended is determined more accurately. can do.

  (5) A means different from the action acquisition means, which is realized by data acquisition means (for example, external data acquisition means 501 and external data conversion means 502) that acquires action history, converts the data, and stores it in the external action history database. The browsing behavior comparing means includes an information recommendation system for analyzing data stored in an external behavior history database (for example, realized by the external behavior history database 503). The information recommendation system configured as described above can determine the content to be recommended more accurately based on more action history.

  The present invention can be effectively applied to an information recommendation system for recommending information with reference to a database and a program for realizing the information recommendation system on a computer.

It is a block diagram which shows the minimum structural example of the information recommendation system by this invention. It is a block diagram which shows the structural example of the information recommendation system of 1st Embodiment. It is a flowchart which shows the example of operation | movement of the information recommendation system in 1st Embodiment. It is a flowchart which shows the example of operation | movement of an action acquisition means. It is a block diagram which shows the structural example of the information recommendation system of 2nd Embodiment. It is explanatory drawing which shows the example of registration of an external action history database. It is explanatory drawing which shows the example of registration of a content database. It is explanatory drawing which shows the example of registration of a browsing history database. It is explanatory drawing which shows the example of registration of an action history database. It is explanatory drawing which shows the example of registration of a browsing action comparison database.

Explanation of symbols

10 Information recommendation system 404 Content determination means

Claims (9)

Content determination means is provided for determining content to be recommended as presentation content based on a combination of a browsing date and time for browsing content and an action date and time based on the browsed content, and determining when to recommend the content. An information recommendation system characterized by that. A browsing behavior comparison database that stores a combination of browsing date and action date and time,
Of the combinations stored in the browsing behavior comparison database, comprising a ratio calculation means for calculating the ratio of the combination of processing objects for each combination,
The information recommendation system according to claim 1, wherein the content determination unit extracts an action date and time for each browsing date based on the ratio calculated by the ratio calculation unit, and determines content related to the extracted action date and time as recommended content. . A content database that stores one or more contents and stores information indicating a location, a period, and a genre of the content target in association with each content;
Referring to the content database, a content content display means for displaying the content of one content selected by the user and storing information indicating the browsing date in the browsing history database;
Action acquisition means for acquiring the user's action and storing information indicating the action date and time in the action history database;
Analyzing the correlation between the browsing history database and the behavior history database, and comprising browsing behavior comparison means for storing a combination of browsing date and behavior date in the browsing behavior comparison database,
The browsing history database stores, for each piece of information indicating a user, a browsing date and time when the user browsed the content and information indicating a location given to the content.
The action history database stores, for each piece of information indicating a user, the action date and time when the user actually acted, and information indicating the place where the user performed,
The browsing behavior comparison database is a combination of the browsing date and behavior date and time, and the combination calculated by the ratio calculation means for the information indicating the user and the information indicating the place, with respect to the entire combination stored in the browsing behavior comparison database. Remember the proportion of
The content determination unit includes a priority determination unit that extracts an action date and time for each browsing date based on the ratio and determines a priority, and the corresponding action based on the priority determined by the priority determination unit The information recommendation system according to claim 2, wherein content related to date and time is determined as recommended content. The browsing behavior comparison means compares the information indicating the location stored in the browsing history database with the information indicating the location stored in the behavior history database, and when both match, the browsing behavior comparison means is based on the browsed content 4. The information recommendation system according to claim 3, wherein the information recommendation system determines and stores the browsing date / time and the behavior date / time in the browsing behavior comparison database using the location as a key. It is a means different from the action acquisition means, including a data acquisition means for acquiring an action history, converting the data and storing it in an external action history database,
The information recommendation system according to claim 3 or 4, wherein the browsing behavior comparison unit analyzes data stored in the external behavior history database by the data acquisition unit. Including a content determination step for determining content to be recommended as a presentation content based on a combination of a browsing date and time for browsing content and an action date and time for acting based on the browsed content, and determining when to recommend the content An information recommendation method characterized by that. A browsing behavior comparison step of storing a combination of the browsing date and behavior date and time in a browsing behavior comparison database;
A ratio calculation step of calculating, for each combination, a ratio occupied by a combination of processing targets among combinations stored in the browsing behavior comparison database;
The information recommendation method according to claim 6, wherein in the content determination step, an action date is extracted for each browsing date based on the ratio calculated in the ratio calculation step, and content related to the extracted action date is determined as recommended content. . On the computer,
Based on the combination of the browsing date and time when the content was browsed and the action date and time when the user acted based on the browsed content, the content to be recommended as the presentation content was determined and the content determination process was performed to determine when to recommend the content Program for recommending information. On the computer,
Browsing behavior comparison processing for storing a combination of browsing date and behavior date in a browsing behavior comparison database;
Of the combinations stored in the browsing behavior comparison database, a ratio calculation process for calculating a ratio of a combination of processing targets for each combination is executed.
A content determination process for executing a process of extracting an action date and time for each browsing date and time based on the ratio calculated in the ratio calculation process, and determining a content related to the extracted action date and time as a recommended content. 8. The information recommendation program according to 8.

Images