Preference learning apparatus, preference learning system, preference learning method, and recording medium

• US 6,741,987 B2
• Filed: 09/26/2001
• Issued: 05/25/2004
• Est. Priority Date: 09/27/2000
• Status: Expired due to Fees

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Ito, Noriko
• Primary Examiner(s): AMSBURY, WAYNE P
• Application Number:
  US09/962,102
• Publication Number:
  US 20020038342A1
• Time in Patent Office: 972 Days
• Field of Search: 707/6
• US Class Current: 707/748
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G09B 5/02 with visual presentation of the material to be studied, e.g. using film strip
• CPC Class Codes: Y10S 707/922 Communications
• CPC Class Codes: Y10S 707/962 Entity-attribute-value
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

A preference learning apparatus includes a content attribute information database, action information database, time information correlation table, area information correlation table, user'"'"'s preference information database, user action detection section, and preference information management section. The preference information management section generates preference analysis data using time zone information, area name, attribute and its weight, and attribute value related to a user'"'"'s preference and updates the weight contained in preference information in the user'"'"'s preference information database, which is specified by time zone information, area name, attribute/attribute value, and user ID, with the weight contained in the generated preference analysis data. A preference learning system, preference learning method, and recording medium are also disclosed.

Claims

1. A preference learning apparatus for detecting an individual user'"'"'s action from each of a plurality of different positions of a portable information terminal to which various kinds of contents are provided through a communication channel and learning the individual user'"'"'s preference on the basis of a detected action log, comprising:

a content attribute information database for storing, for each content, an attribute/attribute value as objects of the learning contained in each of various kinds of contents;

an action information database for storing, for each action, an attribute as an object of the learning estimated from the individual user'"'"'s action and a weight for the attribute;

a time information correlation table for storing a name and time range of a time zone in correspondence with each other;

an area information correlation table for storing each area name and area range in correspondence with each other for each of a plurality of areas which classify position information of the individual user;

an individual user'"'"'s preference information database for storing, for each individual user'"'"'s preference information containing an attribute/attribute value as objects of the learning, a weight for the attribute, a time zone when the weight is valid, and a place where the weight is valid;

user action detection means for detecting the individual user'"'"'s action on the basis of information obtained from the portable information terminal and acquiring detection data containing a user ID indicating the individual user, a action name indicating the action, a content ID indicating a content related to an object of the action, measurement time information at which each of a plurality of different actions has been detected and measurement position information corresponding to each of said plurality of different positions of said portable information terminal at which said plurality of different actions have been detected; and

preference information management means for updating said user'"'"'s preference information database on the basis of preference analysis data obtained by analyzing the user'"'"'s preference on the basis of the detection data output from said user action detection means,

wherein said preference information management means generates the preference analysis data using time zone information acquired from said time information correlation table on the basis of the measurement time contained in the detection data output from said user action detection means, the area name acquired from said area information correlation table on the basis of the position information contained in the detection data, the attribute and weight contained in the action which are acquired from said action information database on the basis of the action name contained in the detection data, and the attribute value acquired from said content attribute information database on the basis of the attribute as an object of the learning and the content ID contained in the detection data, and updates, with a weight contained in the generated preference analysis data, the weight contained in the preference information in the user'"'"'s preference information database, which is specified by the time zone information, area name, and attribute/attribute value as objects of the learning, which are contained in the generated preference analysis data, and the user ID contained in the detection data.

2. An apparatus according to claim 1, further comprising preference information update means for decrementing a value of the weight of the preference information stored in said individual user'"'"'s preference information database when a predetermined time has elapsed from a latest update date/time to re update the value of the weight.

3. An apparatus according to claim 1, wherein the attribute/attribute value as objects of the learning are an attribute/attribute value related to the individual user'"'"'s-preference.

4. A preference learning system having a center for managing various kinds of contents to be provided to an individual user of a portable information terminal which accesses said center through a radio communication channel in accordance with the individual user'"'"'s operation to acquire and display a desired content, and a preference learning apparatus for learning the individual user'"'"'s preference on the basis of information obtained from said portable information terminal in accordance with a time zone and each of a plurality of different positions at which the individual user'"'"'s action takes place, wherein

said preference learning apparatus comprises:

a content attribute information database for storing, for each content, an attribute/attribute value as objects of the learning contained in each of various kinds of contents;

an action information database for storing, for each action, an attribute as an object of the learning estimated from the individual user'"'"'s action and a weight for the attribute;

a time information correlation table for storing a name and time range of a time zone in correspondence with each other;

an area information correlation table for storing each area name and area range in correspondence with each other for each of a plurality of areas which classify position information of the individual user;

an individual user'"'"'s preference information database for storing, for each user'"'"'s preference information containing an attribute/attribute value as objects of the learning, a weight for the attribute, a time zone when the weight is valid, and a place where the weight is valid;

user action detection means for detecting the individual user'"'"'s action on the basis of information obtained from said portable information terminal and acquiring detection data containing a user ID indicating the individual user, a action name indicating the action, a content ID indicating a content related to an object of the action, measurement time information at which each of a plurality of different actions has been detected and measurement position information corresponding to each of said plurality of different positions of said portable information terminal at which said plurality of different actions have been detected; and

preference information management means for generating preference analysis data by analyzing the user'"'"'s preference on the basis of the detection data output from said user action detection means and updating said user'"'"'s preference information database on the basis of the generated preference analysis data, and

said preference information management means generates the preference analysis data using time zone information acquired from said time information correlation table on the basis of the measurement time contained in the detection data output from said user action detection means, the area name acquired from said area information correlation table on the basis of the position information contained in the detection data, the attribute and weight contained in the action acquired from said action information database on the basis of the action name contained in the detection data, and the attribute value acquired from said content attribute information database on the basis of the attribute as an object of the learning and the content ID contained in the detection data, and updates, with the weight contained in the generated preference analysis data, a weight contained in the preference information of the user'"'"'s preference information database, which is specified by the time zone information, area name, and attribute/attribute value as objects of the learning, which are contained in the generated preference analysis data, and the user ID contained in the detection data.

5. A system according to claim 4, wherein said preference learning apparatus is arranged in said center.

6. A system according to claim 4, wherein said preference learning apparatus is arranged in said portable information terminal.

7. A system according to claim 4, wherein

said user action detection means of said preference learning apparatus is arranged in said portable information terminal,

said content attribute information database, action information database, time information correlation table, area information correlation table, user'"'"'s preference information database, and preference information management means of said preference learning apparatus are arranged in said center, and

said preference information management means receives the detection data transmitted from said user action detection means.

8. A system according to claim 4, wherein the attribute/attribute value as objects of the learning are an attribute/attribute value related to the use'"'"'s preference.

9. A preference learning method of detecting an individual user'"'"'s action from each of plaintiffs of different positions of portable information terminal to which various kinds of contents are provided through a communication channel and learning an individual user'"'"'s preference on the basis of a detected action log in a system comprising

a content attribute information database for storing, for each content, an attribute/attribute value as objects of the learning contained in each of various kinds of contents,

an action information database for storing, for each action, an attribute as an object of the learning estimated from the individual user'"'"'s action and a weight for the attribute,

a time information correlation table for storing a name and time range of a time zone in correspondence with each other,

an area information correlation table for storing each area name and area range in correspondence with each other for each of a plurality of areas which classify position information of the individual user, and

an individual user'"'"'s preference information database for storing, for each individual user, user'"'"'s preference information containing an attribute/attribute value as an object of the learning, a weight for the attribute, a time zone when the weight is valid, and a place where the weight is valid, comprising the steps of:

detecting the individual user'"'"'s action on the basis of information obtained from the portable information terminal and acquiring detection data containing a user ID indicating the user, a action name indicating the action, a content ID indicating a content related to an object of the action, measurement time information at which each of a plurality of different actions has been detected and measurement position information corresponding to each of said plurality of different positions of said portable information terminal at which said plurality of different actions have been detected;

generating preference analysis data using time zone information acquired from the time information correlation table on the basis of the measurement time contained in the obtained detection data, the area name acquired from the area information correlation table on the basis of the position information contained in the detection data, the attribute and weight contained in the action which are acquired from the action information database on the basis of the action name contained in the detection data, and the attribute value acquired from the content attribute information database on the basis of the attribute as an object of the learning and the content ID contained in the detection data; and

specifying the preference information in the user'"'"'s preference information database using the time zone information, area name, and attribute/attribute value as objects of the learning, which are contained in the generated preference analysis data, and the user ID contained in the detection data, and updating the weight contained in the specified preference information with the weight contained in the preference analysis data.

10. A method according to claim 9, further comprising the step of decrementing a value of the weight of the preference information stored in the user'"'"'s preference information database when a predetermined time has elapsed from a latest update date/time to re update the value of the weight.

11. A recording medium which stores a program for detecting an individual user'"'"'s action from each of a plurality of different positions of a portable information terminal to which various kinds of contents are provided through a communication channel and learning the individual user'"'"'s preference on the basis of a detected action log in a system comprising

a content attribute information database for storing, for each content, an attribute/attribute value as objects of the learning contained in each of various kinds of contents,

an action information database for storing, for each action, an attribute as an object of the learning estimated from the individual user'"'"'s action and a weight for the attribute,

a time information correlation table for storing a name and time range of a time zone in correspondence with each other,

an area information correlation table for storing each area name and area range in correspondence with each other for each of a plurality of areas which classify position information of the individual user, and

an individual user'"'"'s preference information database for storing, for each individual user, user'"'"'s preference information containing an attribute/attribute value as objects of the learning, a weight for the attribute, a time zone when the weight is valid, and a place where the weight is valid, comprising the steps of:

detecting the individual user'"'"'s action on the basis of information obtained from the portable information terminal and acquiring detection data containing a user ID indicating the individual user, a action name indicating the action, a content ID indicating a content related to an object of the action, measurement time information at which each of a plurality of different actions has been detected and measurement position information corresponding to each of said plurality of different positions of said portable information at which said plurality of different actions have been detected;

generating preference analysis data using time zone information acquired from the time information correlation table on the basis of the measurement time contained in the obtained detection data, the area name acquired from the area information correlation table on the basis of the position information contained in the detection data, the attribute and weight contained in the action which are acquired from the action information database on the basis of the action name contained in the detection data, and the attribute value acquired from the content attribute information database on the basis of the attribute as an object of the learning and the content ID contained in the detection data; and

specifying the preference information in the user'"'"'s preference information database using the time zone information, area name, and attribute/attribute value as objects of the learning, which are contained in the obtained preference analysis data, and the user ID contained in the detection data, and updating the weight contained in the specified preference information with the weight contained in the preference analysis data.

12. A medium according to claim 11, wherein said program further comprises the step of decrementing a value of the weight of the preference information stored in the user'"'"'s preference information database when a predetermined time has elapsed from a latest update date/time to re update the value of the weight.