Information retrieval

• US 20060173830A1
• Filed: 01/19/2006
• Published: 08/03/2006
• Est. Priority Date: 07/23/2003
• Status: Active Grant

• Current Assignee: University College Dublin
• Original Assignee: University College Dublin
• Inventors: Smyth, Barry, Balfe, Evelyn
• Granted Patent: US 7,599,928 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9538 Presentation of query results
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A meta search engine (1) receives a query (q_T) from a user (U), and submits adapted queries (q₁-q_n) to n search engines (S₁-S_n). The results (R₁-R_n) are adapted (R₁′-R_n′) and combined to provide a result set R′. In parallel, the meta search engine (1) accesses a selected hit matrix (H) populated with values for correspondence of queries q₁-q_m with web pages p_i-p_n. These values are the number of “hits” for pervious searches. The pages (R″) retrieved from a row for q_T are combined with the search engine (S₁-S_n) results to provide a final result but which is ranked according to previous search experience. The hit matrix may be dynamically selected according to a user community. The query (q_T) need not exactly match a hit matrix query (q₁-q_n) as rows can be selected according to similarity to the current query q_T with weightings applied.

Claims

1-28. -28. (canceled)

29. A meta search engine for

receiving user search queries,

submitting corresponding queries to search engines and receiving their responses, and

for post-processing the search engine responses to provide a ranked user response list,

wherein the meta search engine ranks the engine responses according to historical data indicating history of correspondence between user queries and response selections;

wherein the meta search engine determines historical selection values for a direct match of a user query; and

wherein the meta search engine determines historical selection values for a plurality of previous queries, which are related to the current query.

30. The meta search engine as claimed in claim 29, wherein the correspondence is stored as a matrix having historical hit numbers for query-selection pairs.

31. The meta search engine as claimed in claim 29, wherein the correspondence is stored as a matrix having historical hit numbers for query-selection pairs, and wherein there is a plurality of matrices, and the meta search engine dynamically selects a matrix.

32. The meta search engine as claimed in claim 29, wherein the correspondence is stored as a matrix having historical hit numbers for query-selection pairs, and wherein there is a plurality of matrices, and the meta search engine dynamically selects a matrix;

and wherein the meta search engine dynamically selects a matrix according to a community to which the user is allocated.

33. The meta search engine as claimed in claim 32, wherein the community is determined according to source of a query.

34. The meta search engine as claimed in claim 32, wherein the community is determined according to source of a query;

and wherein a Web site from which a query originates is used to identify a community.

35. The meta search engine as claimed in claim 31, wherein a log-in procedure of a user is used to determine a hit matrix.

36. The meta search engine as claimed in claim 31, wherein the engine uses a plurality of hit matrices, and firstly presents results arising from use of a primary matrix.

37. The meta search engine as claimed in claim 34, wherein the results are presented together with information about the communities associated with the matrices.

38. The meta search engine as claimed in claim 34, wherein the results are presented together with information about the communities associated with the matrices;

and wherein the results are presented in usage-based clusters.

39. The meta search engine as claimed in claim 31, wherein the engine uses a plurality of matrices and it automatically determines which matrix is likely to be the most relevant.

40. The meta search engine as claimed in claim 39, wherein the engine computes a success score derived from the relative number of selections for the query compared to the total number of selections of the matrix.

41. The meta search engine as claimed in claim 29, wherein weightings are applied to the selection values according to similarity of the associated query to the current query.

42. The meta search engine as claimed in claim 41, wherein the similarity is determined according to extent of query word overlap.

43. The meta search engine as claimed in claim 29, wherein one dimension of a matrix relates to queries, and a weighting is applied to each row or column for that dimension.

44. The meta search engine as claimed in claim 29, wherein the historical data is anonymous as regards particular users, thus ensuring privacy.

45. The meta search engine as claimed in claim 29, wherein the engine modifies the historical data according to a parameter.

46. The meta search engine as claimed in claim 45, wherein the parameter is time, older historical data having less relevance than newer historical data.

47. The meta search engine as claimed in claim 29, wherein the engine combines a determined relevancy score with another score.

48. The meta search engine as claimed in claim 47, wherein the other score is a sponsored listing ranking.

49. The meta search engine as claimed in claim 29, wherein the engine comprises a case based reasoning system in which each query and its historical selection data is a case.

50. The meta search engine as claimed in claim 49, wherein each case comprises a query and k result pairs, each pair having a response identifier and a relevance score.

51. The meta search engine as claimed in claim 50, wherein each case corresponds to a query dimension of a hit matrix.

52. The meta search engine as claimed in claim 31, wherein the meta search engine dynamically selects a matrix according to other communities of searchers.

53. The meta searching method as claimed in claim 31, wherein the search engine uses metrics to identify a matrix which is likely to be the best indicator of results for a query.

54. The computer readable medium comprising software code for implementing operations of a meta search engine of claim 29 when executing on a digital computer.