Method and Apparatus for Identifying, Extracting, Capturing, and Leveraging Expertise and Knowledge

• US 20080104004A1
• Filed: 10/17/2007
• Published: 05/01/2008
• Est. Priority Date: 12/29/2004
• Status: Abandoned Application

• Current Assignee: Monetate, Inc. (StartApp, Inc.)
• Original Assignee: Monetate, Inc. (StartApp, Inc.)
• Inventors: Brave, Scott, Bradshaw, Robert, Minson, Christopher, Jia, Jack
• Application Number:
  US11/874,137
• Publication Number:
  US 20080104004A1
• Time in Patent Office: Days
• US Class Current: 706/45
• CPC Class Codes: G06F 16/901 Indexing; Data structures therefor; Storage structures for retrieval from the web G06F16/951
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

The invention comprises a set of complementary techniques that dramatically improve enterprise search and navigation results. The core of the invention is an expertise or knowledge index, called UseRank that tracks the behavior of website visitors. The expertise-index is designed to focus on the four key discoveries of enterprise attributes: Subject Authority, Work Patterns, Content Freshness, and Group Know-how. The invention produces useful, timely, cross-application, expertise-based search and navigation results. In contrast, traditional Information Retrieval technologies such as inverted index, NLP, or taxonomy tackle the same problem with an opposite set of attributes than what the enterprise needs: Content Population, Word Patterns, Content Existence, and Statistical Trends. Overall, the invention encompasses Baynote Search—a enhancement over existing IR searches, Baynote Guide—a set of community-driven navigations, and Baynote Insights—aggregated views of visitor interests and trends and content gaps.

Claims

1. A computer implemented method for determining knowledge and/or expertise, comprising the steps of:

identifying a target community (or subgroup) of users;

observing usage patterns of Web documents and other online assets by said target community (or subgroup) of users; and

based upon said observed usage patterns, detecting which of said Web documents and other online assets that are most used and useful to said target community (or subgroup) of users.

2. The method of claim 1, further comprising the step of:

based upon said observed usage patterns, automatically determining the topic of said Web documents and other online assets.

3. The method of claim 1, further comprising the steps of:

analyzing a broad range of behaviors exhibited by users while interacting with a Web page or piece of content, said behaviors comprising any of mouse movement, dwell time, scrolling, link usage, searching, repeat visits, and combinations of said actions, as well as absence of said actions; and

determining through said analyzing usefulness of content, contexts associated with said usefulness, and similarities between users.

4. The method of claim 1, further comprising the step of:

transparently executing said method within an existing enterprise application and repository environment based upon a conventional engine provided by a third party;

wherein no user training or adoption of new interfaces is required.

5. The method of claim 1, further comprising the step of:

developing implicit observations that predict user intentions with strong confidence, said observations comprising any of think time, virtual bookmarks, virtual print, and virtual email;

wherein in all cases, users have not necessarily performed a bookmark, print, or email against Web documents.

6. The method of claim 1, further comprising the step of:

providing contextual information derived from said observed usage patterns that a user is not directly asking for via a search query or keyword.

7. The method of claim 1, further comprising the step of:

implementing said method as a wrapper for an existing search mechanism;

wherein when a user issues a search query, said query is handled initially by a dedicated system;

wherein said dedicated system, in turn, forwards said query to said existing search mechanism;

wherein said dedicated system optionally performs one or more searches or related operations against its own internal indexes and databases; and

wherein once results from the searches have been obtained, they are merged together into a single set of results.

8. The method of claim 1, further comprising the steps of:

inferring that results are irrelevant to the user if a user does not perform any action at all against results from a query; and

retaining said inference and using it to influence results of future queries by said user.

9. The method of claim 1, further comprising the step of:

detecting experts by examining a community and individuals who have the ability to discover and collect the most useful Web documents having the most impact.

10. The method of claim 1, further comprising the step of:

providing an inline user interface (UI) rendered using tags comprising a result set derived from a selected view on said observed usage patterns and distilled wisdom from said patterns, including any of a “

most popular,”

“

next step,”

“

similar documents,” and

“

preferred”

tag.

11. The method of claim 1, further comprising the step of:

providing an ordered sequence of search processors in which a first search processor comprises an independent search processor and the second and subsequent search processors act as filters for search processors preceding them.

12. The method of claim 1, further comprising the step of:

providing an explicit bias search processor for recognizing certain queries or query keywords and injecting a fixed set of documents into results for those queries, each with a fixed score.

13. The method of claim 1, further comprising the step of:

providing a popularity search processor comprising an ancillary filter for detecting popular queries and increasing a ranking of documents in results that have historically been selected and used by previous users making the same query.

14. The method of claim 1, further comprising the step of:

dynamically producing a report comparing a ratio of good-to-poor search results for queries that were enhanced using said method to a same ratio for queries that were not enhanced by using said method.

15. The method of claim 1, wherein said method is implemented in an independent and content agnostic system that looks at a location of a Web document and how users interact with that Web document, rather than content of the Web document itself.

16. The method of claim 1, further comprising the steps of:

creating a federation across multiple applications, Websites, and repositories based upon a user'"'"'s pattern of usage of information from and across said multiple applications, Websites, and repositories; and

automatically performing a federated search across said multiple applications, Websites, and repositories, wherein when a query is searched again, information from said multiple applications, Websites, and repositories is recommended.

17. The method of claim 1, further comprising the step of:

performing an augmented search by blending traditional full-text search with preference and activeness information from global, peer, and expert populations.

18. The method of claim 1, further comprising the step of:

returning the n most popular links that have to do with a query term to a user in response to a user query;

wherein if there is no term, then the top n most popular links are returned.

19. The method of claim 18, further comprising the step of:

limiting the set of most popular links and associated assets to those meeting constraints of one or more configured filters.

20. The method of claim 1, further comprising the steps of:

preloading content; and

showing a user a preview of what content is associated with a link.

21. The method of claim 20, further comprising the steps of:

temporarily replacing current content being displayed with previewed content for as long as the preview is active; and

seamlessly returning to the original content view when the preview is deactivated.

22. The method of claim 1, further comprising the step of:

identifying a spectrum of values ranging from short-lived information, to mid-range knowledge, to long-lived wisdom.

23. The method of claim 1, further comprising the step of:

observing usage patterns in a content-agnostic way;

wherein usage patterns can be detected for any content type, including non-HTML content.

24. The method of claim 1, further comprising the steps of:

performing a content gap analysis of use of content, data, and applications by ascertaining what users are actually looking for and what is missing when the user exhibits search or navigation behavior; and

providing a gap analysis report for detected gaps in said content.

25. The method of claim 1, further comprising the steps of:

performing an analysis of use of content over time including what pieces of content are most useful to a community of users at a given time and how that usefulness trends over time; and

providing a content report for visualizing said usage patterns.