Integrated collaborative/content-based filter structure employing selectively shared, content-based profile data to evaluate information entities in a massive information network

• US 6,308,175 B1
• Filed: 11/19/1998
• Issued: 10/23/2001
• Est. Priority Date: 04/04/1996
• Status: Expired due to Term

• Current Assignee: I/P Engine, Inc. (XpresSpa Group, Inc. (f/k/a FORM Holdings Corp.))
• Original Assignee: Lycos Incorporated (Brightcom Group Ltd.)
• Inventors: Lang, Andrew K., Kosak, Donald M.
• Primary Examiner(s): Black, Thomas
• Assistant Examiner(s): Coby, Frantz
• Application Number:
  US09/195,708
• Time in Patent Office: 1,069 Days
• Field of Search: 707/1, 707/2, 707/5, 707/3, 707/10, 707/102
• US Class Current: 707/608
• CPC Class Codes: G06F 16/337 Profile generation, learning or modification
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9536 Search customisation based on social or collaborative filtering
• CPC Class Codes: Y10S 707/959 Network
• CPC Class Codes: Y10S 707/99931 Database or file accessing
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

An information processing system employs an integrated collaborative/content-based information filter structure to find network informons relevant to information desired by an individual user. The filter structure includes a three-level content-based filter which preprocesses informons in a first level to pass only relevant informons, applies community filters in a second level to pass relevant informons to matching communities of users, and applies in a bottom level user filters within each community to pass relevant informons to matching users. Collaborative filtering is integrated with content-based filtering through agent mind melding by which content profile data used in the content-based filtering is formed from a merger of an individual-user or member client data profile and any similar data profiles of other users or other member clients. A threshold level of similarity is required an other-user data profile to be classified as similar for merger with the individual-user data profile.

Claims

1. An information processing system operable in a network having multiple sources of information, the system comprising:

an integrated collaborative/content-based filter structure arranged to process an informon received from the network and to generate profile data representing information in the informon;

the filter structure including a content-based filter which is operative with content profile data representing information desired by an individual user;

a memory system for storing content profile data including the individual user content data profile and content data profiles of other users;

an agent mind melding system for analyzing at least the stored user profile data to identify one or more other user content data profiles similar to the individual user content data profile;

the agent mind melding system merging detected similar content data profiles of other users with the individual user content data profile for operation in the filter structure during filtering of the received informon; and

the filter structure combining at least the informon profile data and the content profile data, including any merged content data profiles of other users, and determining a value of the informon to the individual user.

2. The system of claim 1, wherein a feedback system is provided for collecting and managing a routing of informon rating data to the filter structure from other users who have considered the informon.

3. The system of claim 2 wherein the system is embodied with a computer system having a plurality of microprocessors in a plurality of microprocessor levels, and the microprocessors are interconnected across and between levels to provide facilitated scalability of the system to meet increased loading.

4. The system of claim 1 wherein the agent mind melding system classifies a user profile as being similar to the individual user profile if a threshold of similarity is exceeded in accordance with a measurement based on a selected measure of similarity.

5. The system of claim 4 wherein the stored user profiles analyzed by the agent mind melding system are profiles of users who have considered the informon, and wherein the similar profiles are merged in accordance with a selected data merging proedure.

6. The system of claim 1 wherein the similar profiles are merged in accordance with a selected data merging procedure.

7. An information processing system operable in a network having multiple sources of information, the system comprising:

an integrated collaborative-content-based filter structure having multiple levels and arranged to process an informon received from the network and to generate informon profile data representing information in the informon;

the filter structure including a content-based filter having respective subfilter structures in the respective levels, thereby providing distributed filtering from top to bottom of the filter structure;

the subfilter structure of each level operating in response to content profile data applicable to such level and reflecting information desired by an individual user to provide progressively narrower information filtering relative to desired user information from top to bottom of the filter structure;

a memory system for storing content profile data including first content profile data for operation in at least a lowest of the subfilter structures and second content profile data for operation in a next-to-the-lowest of the subfilter structures;

an agent mind melding system for analyzing at least one of the first and second stored content profile data to identify one or more other content profiles associated with other users at the associated filter level and similar to a content profile associated with the individual user at the associated filter level;

the agent mind melding system merging detected similar content profiles of other users with the individual user content profile at the associated filter level; and

the filter structure combining at least the informon profile data and the individual-user associated content profile with any merged user content profiles, and determining a value of the informon to the individual user.

8. The system of claim 7 wherein the filter structure receives evaluative feedback data from the individual user in relation to the informon and processes the evaluative feedback data to update the individual-user content profile data for better prediction of values of informons subsequently processed for the individual user.

9. The system of claim 8 wherein a feedback system is provided for collecting and managing feedback routing of informon rating data from users who are other than the individual user and who have considered the informon;

and

the filter structure further combines any feedback rating data received from the feedback system in determining the informon value.

10. The system of claim 9 wherein the lowest level of the filter structure is a user level where user content profiles are applied in user filters, wherein the next-to-the-lowest level of the filter structure is a community level where community content profiles each represent user content profiles of users within such community and are applied in associated community filters, and wherein a highest level of the filter structure is a preprocessing level with a content profile representing the community content profiles and being applied in a preprocessing filter.

11. The system of claim 10 wherein the filters in each level are embodied with one or more microprocessors interconnected within levels and between levels to form an overall computer system for filter processing of informons, thereby providing facilitated scalability of the system with expansion of network user load.

12. An information processing system operable in a network having multiple sources of information, the system comprising:

means for processing an informon received from the network with integrated collaborative/content-based filtering and generating profile data representing information in the informon;

the processing means including a content-based filter which is operative with content profile data representing information desired by an individual user;

means for storing content profile data including the individual user content data profile and content data profiles of other users;

means for melding agent minds including means for analyzing at least the stored user profile data to identify one or more other user content data profiles similar to the individual user content data profile;

the agent mind melding means merging detected similar content data profiles of other users with the individual user content data profile for operation in the filter during filtering of the received informon; and

the processing means combining at least the informon profile data and the content profile data, including any merged content data profiles of other users, and determining a value of the informon to the individual user.

13. A method for operating an information processing system in a network having multiple sources of information, the steps of the method comprising:

processing an informon received from the network with multi-level, integrated collaborative/content-based filtering and generating informon profile data representing information in the informon;

the processing step including operating a content-based filter having respective subfilter structures in the respective levels, thereby providing distributed filtering from top to bottom of the filter structure;

operating the subfilter structure of each level in response to content profile data applicable to such level and reflecting information desired by an individual user to provide progressively narrower information filtering relative to desired user information from top to bottom of the filter;

storing content profile data including first content profile data for operation in at least a lowest of the subfiter structures and second content profile data for operation in a next-to-the-lowest of the subfilter structures;

melding agent minds including analyzing at least one of the first and second stored content profile data to identify one or more other content profiles associated with other users at the associated filter level and being similar to a content profile associated with the individual user at the associated filter level;

the melding step further including merging detected similar content profiles of other users with the individual user content profile at the associated filter level; and

combining at least the informon profile data and the individual-user associated content profile with any merged user content profiles, and determining a value of the informon to the individual user.

14. A method for operating an information processing system in a network having multiple sources of information, the steps of the method comprising:

processing an informon received from the network with an integrated collaborative/content-based filter structure and generating profile data representing information in the informon;

the processing step including operating a content-based filter which is operative with content profile data representing information desired by an individual user;

storing content profile data including the individual user content data profile and content data profiles of other users;

melding agent minds including analyzing at least the stored user profile data to identify one or more other user content data profiles similar to the individual user content data profile;

the melding step further including merging detected similar content data profiles of other users with the individual user content data profile for operation in the filter structure during filtering of the received informon; and

combining at least the informon profile data and the content profile data, including any merged content data profiles of other users, and determining a value of the informon to the individual user.

15. An information processing system operable in a network having multiple sources of information, the system comprising:

means for processing an informon received from the network with a multi-level, integrated collaborative/content-based filter structure and for generating informon profile data representing information in the informon;

the processing means including a content-based filter having respective subfilter structures in the respective levels, thereby providing distributed filtering from top to bottom of the filter structure;

the subfilter structure of each level operating in response to content profile data applicable to such level and reflecting information desired by an individual user to provide progressively narrower information filtering relative to desired user information from top to bottom of the filter structure;

means for storing content profile data including first content profile data for operation in at least a lowest of the subfilter structures and second content profile data for operation in a next-to-the-lowest of the subfilter structures;

means for melding agent minds including analyzing at least one of the first and second stored content profile data to identify one or more other content profiles associated with other users at the associated filter level and similar to a content profile associated with the individual user at the associated filter level;

the agent mind melding means merging detected similar content profiles of other users with the individual user content profile at the associated filter level;

the processing means combining at least the informon profile data and the individual-user associated content profile with any merged user content profiles, and determining a value of the informon to the individual user.