Method and apparatus for a user interest topology based on seeded user interest modeling

Patent metadata

Field	Value
Publication number	US-9665648-B2
Application number	US-201013637001-A
Country	US
Kind code	B2
Filing date	Mar 29, 2010
Priority date	Mar 29, 2010
Publication date	May 30, 2017
Grant date	May 30, 2017

Abstract

Methods and apparatuses are provided for user interest modeling. A method may include receiving an input from a user for specifying one or more topics from among a predetermined hierarchy of topics and subtopics. The method may additionally include retrieving one or more documents associated with the user and extracting language tokens from the documents based, at least in part, on the specified topics. Corresponding apparatuses are also provided.

First claim

What is claimed is:

1. A method comprising:
providing, on a user interface associated with an apparatus, a presentation of a menu of one or more topics from among a predetermined hierarchy of topics and subtopics to a user;
receiving, via the user interface by the apparatus, a first input from the user specifying at least one of the one or more topics from among the predetermined hierarchy of topics and subtopics;
retrieving, by the apparatus, one or more documents associated with the user;
extracting, by the apparatus, noun tokens from the documents based, at least in part, on the specified topics;
performing, by the apparatus, histogram processing of the extracted noun tokens to provide a list of pertinent tokens;
comparing, by the apparatus, topic input information received from the user with the list of pertinent tokens to determine if there is a corresponding match;
generating, by the apparatus, a topology of the matching tokens according to a probabilistic model, wherein the topology matches the matching pertinent tokens with the topics and subtopics of the hierarchy;
providing, on the user interface by the apparatus, a presentation of top-level topics as topic icons on a tool bar on the user interface based, at least in part, on the topology of matching tokens; and
in response to a second input from the user specifying one of the topic icons, providing, on the user interface by the apparatus, a presentation of one or more search results customized for the user using machine learning based on the topology.

2. A method of claim 1 , wherein the extracting of the noun tokens comprises:
identifying substantially all noun tokens in the documents;
selecting one of the identified noun tokens based, at least in part, on whether the one identified noun token is semantically related to the specified topics, wherein the extracted noun tokens include the selected one identified noun token;
providing, on the user interface by the apparatus, a presentation of topics determined based on the topology; and
receiving, via the user interface by the apparatus, a third input from the user specifying a subset of the topics, wherein the topic icons correspond to the subset of the topics.

3. A method of claim 1 , further comprising:
associating one of the topics and subtopics of the hierarchy with one or more reference documents; and
extracting a set of reference noun tokens from the reference documents,
wherein the generating of the topology comprises, at least in part, matching the extracted noun tokens with the set of reference noun tokens and associating the extracted noun tokens with the one topic or subtopic based on the matching.

4. A method of claim 3 , further comprising:
translating the topics and subtopics to another language;
associating another one of the topics and subtopics of the hierarchy with one or more reference documents in the another language; and
extracting another set of reference noun tokens specific to the another language;
wherein the generating of the topology further comprises, at least in part, matching the extracted noun tokens with the another set of reference noun tokens in the another language.

5. A method of claim 1 , further comprising:
calculating an entropy associated with the topology, wherein the entropy is a measure of a variety of the topics and subtopics matched to the matching pertinent tokens in the topology, wherein the apparatus is a network node, and the one or more search results are provided to the user interface via a web portal.

6. A method of claim 1 , wherein the topics and subtopics comprise user characteristics, interests, preferences, or a combination thereof, the method further comprising:
causing, at least in part, assignment of the topology to a user profile associated with an application, a service, or combination thereof.

7. A method of claim 1 , wherein the probabilistic model is a latent Dirichlet allocation that is aware of the topics and subtopics of the hierarchy.

8. A method of claim 1 , further comprising:
specifying a maximum number of the topics and subtopics of the hierarchy to match, wherein the topology is generated based, at least in part, on the maximum number;
customizing one or more applications, one or more services, or a combination thereof, to the user based, at least in part, on the topology; and
transmitting the one or more applications, the one or more services, or a combination thereof to the user.

9. A method of claim 1 , further comprising:
specifying a threshold probability for matching the extracted noun tokens with the topics and subtopics of the hierarchy,
wherein the topology is generated based, at least in part, on the threshold probability.

10. An apparatus comprising:
at least one processor; and
at least one memory including computer program code,
the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus to perform at least the following,
provide, on a user interface associated with the apparatus, a presentation of a menu of one or more topics from among a predetermined hierarchy of topics and subtopics to a user;
receive, via the user interface, a first input from the user specifying at least one of the one or more topics from among the predetermined hierarchy of topics and subtopics;
retrieve one or more documents associated with the user;
extract noun tokens from the documents based, at least in part, on the specified topics;
perform histogram processing of the extracted noun tokens to provide a list of pertinent tokens;
compare topic input information received from the user with the list of pertinent tokens to determine if there is a corresponding match;
generate a topology of the matching tokens according to a probabilistic model, wherein the topology matches the matching pertinent tokens with the topics and subtopics of the hierarchy;
provide, on the user interface, a presentation of top-level topics as topic icons on a tool bar on the user interface based, at least in part, on the topology of matching tokens; and
in response to a second input from the user specifying one of the topic icons, provide, on the user interface, a presentation of one or more search results customized for the user using machine learning based on the topology.

11. An apparatus of claim 10 , wherein the extracting of the noun tokens further causes the apparatus to:
identify substantially all noun tokens in the documents;
select one of the identified noun tokens based, at least in part, on whether the one identified noun token is semantically related to the specified topics, wherein the extracted noun tokens include the selected one identified noun token;
provide, on the user interface, a presentation of topics determined based on the topology; and
receive, via the user interface, a third input from the user specifying a subset of the topics, wherein the topic icons correspond to the subset of the topics.

12. An apparatus of claim 10 , wherein the apparatus is further caused to:
associate one of the topics and subtopics of the hierarchy with one or more reference documents; and
extract a set of reference noun tokens from the reference documents,
wherein the generating of the topology comprises, at least in part, matching the extracted noun tokens with the set of reference noun tokens and associating the extracted noun tokens with the one topic or subtopic based on the matching.

13. An apparatus of claim 12 , wherein the apparatus is further caused to:
translate the topics and subtopics to another language;
associate another one of the topics and subtopics of the hi

Assignees

• Sathish Sailesh
• Tian Jilei
• Hu Rile
• Nokia Technologies Oy

Inventors

• Sathish Sailesh
• Tian Jilei
• Hu Rile

Classifications

• G06F17/30867 Primary

  Physics · mapped topic

• G06F17/30734

  Physics · mapped topic

• G06F16/9535 Primary

  Search customisation based on user profiles and personalisation · CPC title

• G06F16/367

  Ontology · CPC title