Categorizing network resources and extracting user interests from network activity

• US 9,311,386 B1
• Filed: 04/03/2013
• Issued: 04/12/2016
• Est. Priority Date: 04/03/2013
• Status: Active Grant

• Current Assignee: The Boeing Co.
• Original Assignee: Narus, Inc. (Gen Digital Inc.)
• Inventors: Song, Han See, Baldi, Mario, Nucci, Antonio, Ma, Qiang
• Primary Examiner(s): Mofiz, Apu
• Assistant Examiner(s): Daye, Chelcie
• Application Number:
  US13/855,872
• Time in Patent Office: 1,105 Days
• Field of Search: 707/740, 707/738
• US Class Current: 1/1
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/3323 using document space presentation or visualization, e.g. category, hierarchy or range presentation and selection
• CPC Class Codes: G06F 16/337 Profile generation, learning or modification
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: G06F 16/38 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9538 Presentation of query results

Abstract

A method for network resource classification and identifying user interests based on the classification. The method uses a provided hierarchy of categories for classifying network resources, wherein each category is assigned a text item describing the category and the method includes obtaining resource description data collections corresponding to the network resources, and generating, using a semantic correlation algorithm, a category score vector of a network resource by comparing the resource description data collection to the text item assigned to each category in the hierarchy of categories, wherein the category score vector comprises a category score for each category in the hierarchy of categories, wherein the category score is determined based on at least a semantic correlation measure between the resource description data collection and the text item assigned to a corresponding category, wherein the plurality of network resources are classified based at least on the category score.

Claims

1. A method for network resource classification, comprising:

obtaining a hierarchy of categories for classifying a plurality of network resources, where each category is assigned a text item describing the category;

obtaining a plurality of resource description data collections corresponding to the plurality of network resources, wherein the plurality of resource description data collections comprise a first resource description data collection corresponding to a first network resource of the plurality of network resources;

generating, by a computer processor and using a semantic correlation algorithm, a first category score vector of the first network resource by comparing the first resource description data collection to the text item assigned to each category in the hierarchy of categories, wherein the first category score vector comprises a category score for each category in the hierarchy of categories, wherein the category score is determined based on at least a semantic correlation measure between the first resource description data collection and the text item assigned to a corresponding category,

wherein the plurality of network resources are classified based at least on the category score;

analyzing a network trace associated with a user to identify the plurality of network resources accessed by the user;

generating, based on a pre-determined criterion, a relationship graph comprising;

a plurality of nodes representing the plurality of network resources, and

a plurality of edges representing a measure of cross-references between the plurality of resource description data collections; and

adjusting, based on the relationship graph, the first category score vector to generate a first adjusted category score vector using at least another category score vector of another network resource of the plurality of network resources.

2. The method of claim 1,

wherein the category score is determined further based on at least another semantic correlation measure between the first resource description data collection and another text item assigned to a neighbor category, wherein the neighbor category is within a pre-determined number of hops in the hierarchy of categories from the corresponding category assigned the text item.

3. The method of claim 1, wherein the first network resource comprises a first network client application, wherein obtaining the plurality of resource description data collections comprises:

retrieving a first description webpage from an online source of the first network client application;

obtaining an identifier of a second network client application based on the first description webpage; and

retrieving a second description webpage from the online source of the second network client application,

wherein the first resource description data collection comprises the first description webpage and the second description webpage.

4. The method of claim 1, wherein the first network resource comprises a first webpage, wherein obtaining the plurality of resource description data collections comprises:

obtaining an URL of a second webpage based on the first webpage,

wherein the first resource description data collection comprises the first webpage and the second webpage.

5. The method of claim 1, wherein the first network resource comprises a network hostname, wherein obtaining the plurality of resource description data collections comprises:

obtaining a plurality of search results using the hostname as a search phrase of a pre-determined search engine,

wherein the first resource description data collection comprises the plurality of search results.

6. The method of claim 1, further comprising:

inferring an interest level of the user among the hierarchy of categories based on the first adjusted category score vector.

7. The method of claim 1, further comprising:

analyzing the first category score vector to determine a first score differentiation measure representing variations among category scores in the first category score vector; and

analyzing the first adjusted category score vector to determine a first adjusted score differentiation measure representing variations among adjusted category scores in the first adjusted category score vector;

wherein adjusting the first category score vector is in response to the first score differentiation measure failing to meet a first pre-determined criterion, and

wherein the interest level of the user is inferred in response to the first adjusted score differentiation measure meeting the first pre-determined criterion.

8. The method of claim 1, wherein adjusting the first category score vector comprises:

selecting, based on the relationship graph and corresponding to a second network resource, a node from the plurality of nodes, wherein the measure of cross-references between the first resource description data collection and a second resource description data collection of the second network resource meets a second pre-determined criterion;

generating, using the semantic correlation algorithm, a second category score vector of the second network resource; and

aggregating, based on a pre-determined formula, the first category score vector and the second category score vector to generate the first adjusted category score vector.

9. The method of claim 1,

wherein the second network resource is separate from the plurality of network resources and has not been used by the user, and

wherein the plurality of edges further represents the measure of cross-references between the second resource description data record and the plurality of resource description data collections.

10. A system for network resource classification, comprising:

a computer processor;

an inference input module executing on the computer processor and configured to;

obtain a hierarchy of categories for classifying a plurality of network resources, where each category is assigned a text item describing the category;

obtain a plurality of resource description data collections corresponding to the plurality of network resources, wherein the plurality of resource description data collections comprise a first resource description data collection corresponding to a first network resource of the plurality of network resources;

analyze a network trace associated with a user to identify the plurality of network resources accessed by the user;

generate, based on a pre-determined criterion, a relationship graph comprising;

a plurality of nodes representing the plurality of network resources,

a plurality of edges representing a measure of cross-references between the plurality of resource description data collections;

a category score generator executing on the computer processor and configured to;

generate, using a semantic correlation algorithm, a first category score vector of the first network resource by comparing the first resource description data collection to the text item assigned to each category in the hierarchy of categories, wherein the first category score vector comprises a category score for each category in the hierarchy of categories, wherein the category score is determined based on at least a semantic correlation measure between the first resource description data collection and the text item assigned to a corresponding category;

adjust, based on the relationship graph, the first category score vector to generate the first adjusted category score vector using at least another category score vector of another network resource of the plurality of network resources; and

a repository configured to store the plurality of resource description data collections and the relationship graph,

wherein the plurality of network resources are classified based at least on the category score.

11. The system of claim 10,

wherein the category score is determined further based on at least another semantic correlation measure between the first resource description data collection and another text item assigned to a neighbor category, wherein the neighbor category is within a pre-determined number of hops in the hierarchy of categories from the corresponding category assigned the text item.

12. The system of claim 10, wherein the first network resource comprises a first network client application, wherein obtaining the plurality of resource description data collections comprises:

retrieving a first description webpage from an online source of the first network client application;

obtaining an identifier of a second network client application based on the first description webpage; and

retrieving a second description webpage from the online source of the second network client application,

wherein the first resource description data collection comprises the first description webpage and the second description webpage.

13. The system of claim 10, wherein the first network resource comprises a first webpage, wherein obtaining the plurality of resource description data collections comprises:

obtaining an URL of a second webpage based on the first webpage,

wherein the first resource description data collection comprises the first webpage and the second webpage.

14. The system of claim 10, wherein the first network resource comprises a network hostname, wherein obtaining the plurality of resource description data collections comprises:

obtaining a plurality of search results using the hostname as a search phrase of a pre-determined search engine,

wherein the first resource description data collection comprises the plurality of search results.

15. The system of claim 10,

wherein the inference input module is further configured to:

infer an interest level of the user among the hierarchy of categories based on a first adjusted category score vector.

16. The system of claim 15, further comprises an inference controller configured to:

analyze the first category score vector to determine a first score differentiation measure representing variations among category scores in the first category score vector; and

analyze the first adjusted category score vector to determine a first adjusted score differentiation measure representing variations among adjusted category scores in the first adjusted category score vector;

wherein adjusting the first category score vector is in response to the first score differentiation measure failing to meet a first pre-determined criterion, and

wherein the interest level of the user is inferred in response to the first adjusted score differentiation measure meeting the first pre-determined criterion.

17. The system of claim 15, wherein adjusting the first category score vector comprises:

selecting, based on the relationship graph and corresponding to a second network resource, a node from the plurality of nodes, wherein the measure of cross-references between the first resource description data collection and a second resource description data collection of the second network resource meets a second pre-determined criterion;

generating, using the semantic correlation algorithm, a second category score vector of the second network resource; and

aggregating, based on a pre-determined formula, the first category score vector and the second category score vector to generate the first adjusted category score vector.

18. The system of claim 15,

wherein the second network resource is separate from the plurality of network resources and has not been used by the user, and

wherein the plurality of edges further represents the measure of cross-references between the second resource description data record and the plurality of resource description data collections.

19. A non-transitory computer readable medium embodying instructions for network resource classification, the instructions when executed by a processor comprising functionality for:

obtaining a hierarchy of categories for classifying a plurality of network resources, where each category is assigned a text item describing the category;

obtaining a plurality of resource description data collections corresponding to the plurality of network resources, wherein the plurality of resource description data collections comprise a first resource description data collection corresponding to a first network resource of the plurality of network resources;

generating, using a semantic correlation algorithm, a first category score vector of the first network resource by comparing the first resource description data collection to the text item assigned to each category in the hierarchy of categories, wherein the first category score vector comprises a category score for each category in the hierarchy of categories, wherein the category score is determined based on at least a semantic correlation measure between the first resource description data collection and the text item assigned to a corresponding category,

wherein the plurality of network resources are classified based at least on the category score

analyzing a network trace associated with a user to identify the plurality of network resources accessed by the user;

generating, based on a pre-determined criterion, a relationship graph comprising;

a plurality of nodes representing the plurality of network resources, and

a plurality of edges representing a measure of cross-references between the plurality of resource description data collections; and

adjusting, based on the relationship graph, the first category score vector to generate a first adjusted category score vector using at least another category score vector of another network resource of the plurality of network resources.

20. The non-transitory computer readable medium of claim 19,

wherein the category score is determined further based on at least another semantic correlation measure between the first resource description data collection and another text item assigned to a neighbor category, wherein the neighbor category is within a pre-determined number of hops in the hierarchy of categories from the corresponding category assigned the text item.

21. The non-transitory computer readable medium of claim 19, wherein the first network resource comprises a first network client application, wherein obtaining the plurality of resource description data collections comprises:

retrieving a first description webpage from an online source of the first network client application;

obtaining an identifier of a second network client application based on the first description webpage; and

retrieving a second description webpage from the online source of the second network client application,

wherein the first resource description data collection comprises the first description webpage and the second description webpage.

22. The non-transitory computer readable medium of claim 19, wherein the first network resource comprises a first webpage, wherein obtaining the plurality of resource description data collections comprises:

obtaining an URL of a second webpage based on the first webpage,

wherein the first resource description data collection comprises the first webpage and the second webpage.

23. The non-transitory computer readable medium of claim 19, wherein the first network resource comprises a network hostname, wherein obtaining the plurality of resource description data collections comprises:

obtaining a plurality of search results using the hostname as a search phrase of a pre-determined search engine,

wherein the first resource description data collection comprises the plurality of search results.

24. The non-transitory computer readable medium of claim 19, wherein an interest level of the user among the hierarchy of categories is inferred based on the first adjusted category score vector.

25. The non-transitory computer readable medium of claim 24, the instructions when executed by a processor further comprising functionality for:

analyzing the first category score vector to determine a first score differentiation measure representing variations among category scores in the first category score vector; and

analyzing the first adjusted category score vector to determine a first adjusted score differentiation measure representing variations among adjusted category scores in the first adjusted category score vector;

wherein adjusting the first category score vector is in response to the first score differentiation measure failing to meet a first pre-determined criterion, and

wherein the interest level of the user is inferred in response to the first adjusted score differentiation measure meeting the first pre-determined criterion.

26. The non-transitory computer readable medium of claim 24, wherein adjusting the first category score vector comprises:

selecting, based on the relationship graph and corresponding to a second network resource, a node from the plurality of nodes, wherein the measure of cross-references between the first resource description data collection and a second resource description data collection of the second network resource meets a second pre-determined criterion;

generating, using the semantic correlation algorithm, a second category score vector of the second network resource; and

aggregating, based on a pre-determined formula, the first category score vector and the second category score vector to generate the first adjusted category score vector.

27. The non-transitory computer readable medium of claim 24,

wherein the second network resource is separate from the plurality of network resources and has not been used by the user, and

wherein the plurality of edges further represents the measure of cross-references between the second resource description data record and the plurality of resource description data collections.