Natural language meta-search system and method

• US 6,078,914 A
• Filed: 12/09/1996
• Issued: 06/20/2000
• Est. Priority Date: 12/09/1996
• Status: Expired due to Term

• Current Assignee: Open Text Corporation
• Original Assignee: Open Text Corporation
• Inventors: Redfern, Darren M.
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): Min, Donald
• Application Number:
  US08/769,929
• Time in Patent Office: 1,289 Days
• Field of Search: 707/2, 707/3, 707/4, 707/5, 707/1
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3334 Selection or weighting of terms from queries, including natural language queries
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• 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 707/99934 Query formulation, input preparation, or translation
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A meta search system accepts natural language queries which are parsed to extract relevant content, this relevant content being formed into queries suitable for each of a selected number of search engines and being transmitted thereto. The results from the search engines are received and examined and a selected number of the information sources represented therein are obtained. These obtained information sources are then examined to rank their relevance to the extracted relevant content and the portions of interest in each of these ranked information sources are determined. The determined portions are output to the user in ranked order, having first been processed to clean up the portions to include valid formatting and complete paragraphs and/or sentences.

Claims

1. A method of locating information in at least one information source using at least one search engine, comprising the steps of:

(i) accepting a natural language query describing desired information;

(ii) parsing said natural language query to extract terms relevant to said desired information;

(iii) creating search data comprising at least two search candidates from said extracted terms in a form appropriate to each of said at least one search engine, and transferring said created search data to said each of said at least one search engine to initiating a search;

(iv) receiving search results comprising at least one list of information sources from said each of said at least one search engine, and removing redundancies from said at least one list of information sources to obtain a reduced list of information sources;

(v) retrieving complete copies of each information source in said reduced list;

(vi) examining each said retrieved complete copy relative to said at least two search candidates to determine a match ranking therefor by;

(a) arranging each said complete copy into segments, each segment defining the contents of said document between at least three consecutive matches between said complete copy and any of said at least two search candidates;

(b) examining each segment in said complete copy to determine a segment score comprising a score for each match between the contents of said complete copy and each search candidate, and weighting said segment score with respect to the length of said segment;

(c) selecting at least two segments of said complete copy with the highest weighted segment scores from step (b);

(d) for each selected segment, augmenting the segment to include the contents of said complete copy between the selected segment and an adjacent match and performing step (b) for each augmented segment to obtain an updated segment score;

(e) while said updated segment score for an augmented segment is greater than said segment store, performing step (d);

(f) selecting said augmented segment with the highest updated segment score from each said complete copy; and

(g) ranking the selected augmented segments for each said complete copy according to said updated segment scores;

(vii) selecting at least the highest ranked selected augmented segment for display to the user, and editing each said at least highest ranked selected segment to form a complete segment by examining the beginning and end of said segment and adding or removing adjacent content of said complete copy to form a substantially grammatically correct segment; and

(viii) providing said each said substantially grammatically correct segment to said user.

2. The method as defined in claim 1 wherein step (vii) further comprises the step of removing predefined formatting information from said segment.

3. The method as claimed in claim 2 wherein said formatting information comprises HTML tags.

4. The method of claim 1 wherein step (vii), said complete segments comprise at least one complete sentence.

5. The method of claim 1 where in step (vii), said complete segments comprise at least one complete paragraph.

6. The method of claim 1 wherein said at least one information source comprises an HTML formatted document.

7. The method of claim 6 wherein said at least one information source is accessible via a telecommunications network.

8. The method of claim 7 wherein said at least one search engine is accessible via a telecommunications network.

9. The method of claim 6 wherein at least first and second retrieved complete copies are obtained, said first complete retrieved copy being obtained from a first location connected to said telecommunications network and said second complete retrieved copy being obtained from a second location connected to said telecommunications network.

10. The method of claim 6 wherein at least first and second search engines are employed, said first search engine being located at a first location connected to said telecommunications network and said second search engine being located at a second location connected to said telecommunications network.

11. The method of claim 6 wherein said at least one information source is located on a computer network.

12. The method of claim 11, further comprising a plurality of information sources, each information source being located on said computer network.

13. The method of claim 11 wherein said at least one search engine is located on said computer network.