Method and apparatus for deriving logical relations from linguistic relations with multiple relevance ranking strategies for information retrieval

• US 6,829,605 B2
• Filed: 05/24/2001
• Issued: 12/07/2004
• Est. Priority Date: 05/24/2001
• Status: Active Grant

• Current Assignee: Zhigu Holdings Limited
• Original Assignee: Microsoft Corporation
• Inventors: Azzam, Saliha
• Primary Examiner(s): Kindred, Alford
• Assistant Examiner(s): Wong, Leslie
• Application Number:
  US09/865,032
• Publication Number:
  US 20020178152A1
• Time in Patent Office: 1,293 Days
• Field of Search: 707/1-7, 707/102, 707/104.1
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: Y10S 707/922 Communications
• 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
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing

Abstract

A method and apparatus are provided for improving the precision of information retrieval systems that use logical form searching techniques. Under one embodiment of the invention, several logical form triples are produced from the user'"'"'s query and are combined together by restrictive logical operators to generate a compound logical form query. A search is then performed to find documents that meet the requirements set by the compound logical form query. In other embodiments, results generated by a logical form search are intersected with results from a word search to form a more precise set of results. In further embodiments of the invention, three pairs of search results are intersected with each other to form three sets of final results. These final results are then ranked based on the techniques used to form their constituent result pairs.

Claims

1. A method of retrieving information from a document collection, the method comprising:

generating a compound logical form query from a user query, the compound logical form query having at least two logical form triples connected by a restrictive operator;

applying the compound logical form query to a logical form triple index of a document collection to form a set of logic-based search results;

generating a word query based on words in the user query;

applying the word query to a word index of the document collection to form a set of word-based search results;

intersecting the logic-based search results with the word-based search results to form a set of filtered compound results;

generating a standard logical form triple query by connecting a set of logical form triples with non-restrictive operators;

applying the standard logical form triple query to the logical form triple index of the document collection to form a second set of logic-based search results;

intersecting the word-based search results with the second set of logic-based search results to form a set of filtered standard results; and

returning the filtered compound results and the filtered standard results to the user with the filtered compound results ranked higher than the filtered standard results.

2. The method of claim 1 wherein generating a compound logical form query comprises:

determining the restrictiveness of a set of logical form triples identified from the user query; and

combining logical form triples so as to limit the restrictiveness of the compound logical form query formed from the set of logical form triples.

3. The method of claim 1 wherein generating a word query comprises identifying important words in the user query and placing the important words in the word query.

4. The method of claim 1 further comprising:

generating a second word query based on words in the user query;

applying the second word query to the word index to form a second set of word-based search results; and

intersecting the second set of word-based search results with the filtered standard results to form further filtered standard results.

5. The method of claim 4 further comprising returning the filtered compound results, the further filtered standard results and the filtered standard results to the user with the filtered compound results ranked above the further filtered standard results and the further filtered standard results ranked above the filtered standard results.

6. A computer-readable medium having computer-executable instructions for performing information retrieval steps comprising:

performing a logic-based search of a document collection using logical form triples created from a user query to create logic-based search results, performing the logic-based search comprising forming a compound logical form query by identifying at least two logical form triples from the user query and connecting at least two of the logical form triples with a restrictive operator, and using the compound logical form query to search the document collection;

performing a word-based search of the document collection using words taken from the user query to create word-based search results;

intersecting the logic-based search results and the word-based search results to form filtered search results;

performing a standard logic-based search of the document collection by identifying at least two logical form triples from the user query, connecting each of the identified logical form triples together using non-restrictive operators to form a standard logical form triple query, and searching the document collection using the standard logical form triple query to form standard logic-based search results;

intersecting the standard logic-based search results with the word-based search results to form filtered standard search results; and

returning the filtered search results and the filtered standard search results to the user with the filtered search results ranked higher than the filtered standard search results.

7. The computer-readable medium of claim 6 wherein performing a word-based search comprises identifying important words in the user query and using only the important words in the word-based search.

8. The computer-readable medium of claim 6 wherein the information retrieval steps further comprise:

performing a second word-based search of the document collection to form second word-based search results; and

intersecting the second word-based search results with the filtered standard search results to form further filtered standard search results.

9. The computer-readable medium of claim 8 wherein the information retrieval steps further comprise returning the filtered search results, the filtered standard search results, and the further filtered standard search results to the user with the filtered search results ranked higher than the further filtered standard search results and the further filtered standard search results ranked higher than the filtered standard search results.