SEARCHING A DATABASE OF LISTINGS

• US 20080281806A1
• Filed: 05/10/2007
• Published: 11/13/2008
• Est. Priority Date: 05/10/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Ju, Yun-Cheng, Yu, Dong, Acero, Alejandro, Wang, Ye-Yi, Zweig, Geoffrey G.
• Granted Patent: US 9,218,412 B2
• Time in Patent Office: Days
• US Class Current: 707/5
• CPC Class Codes: G06F 16/3334 Selection or weighting of terms from queries, including natural language queries
• CPC Class Codes: G06F 16/3347 using vector based model
• CPC Class Codes: G06F 3/0641 De-duplication techniques
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
• CPC Class Codes: G10L 15/197 Probabilistic grammars, e.g. word n-grams

Abstract

A database having listings rather than long documents is searched using a term frequency-inverse document frequency (Tf/Idf) algorithm.

Claims

1. A computer system for searching a database, the system comprising:

a database having a plurality of listings;

a query recognizer configured to receive a query from a user and provide recognized content indicative of the query;

a search module configured to receive the recognized content, search the database using a term frequency-inverse document frequency (Tf/Idf) algorithm based on the recognized content and provide at least one listing associated with the query.

2. The computer system of claim 1 wherein one of the query recognizer or the search module is configured to modify the recognized content to change duplicative terms to unique terms, and wherein the database includes listings where otherwise duplicative terms have been changed to unique terms.

3. The computer system of claim 1 wherein the database is a structured database having a plurality of fields and wherein a plurality of listings have terms in each of their corresponding fields, and wherein the search module is configured to search each of the fields based on the recognized content and provide a plurality of listings as search results, and wherein the search module is configured to analyze the search results to discard at least one listing based on the analysis.

4. The computer system of claim 1 wherein the database is a structured database comprising a plurality of fields and wherein a plurality of listings have terms in each of their corresponding fields, wherein the search module is configured to search each of the fields based on the recognized content and provide a plurality of listings as search results, and wherein the search module is configured to analyze the recognized content to discard a term in the recognized content to form a redacted recognized content and search the results to discard at least one listing based on the analysis.

5. The computer system of claim 1 wherein the database is a structured database comprising a plurality of fields and wherein a plurality of listings have terms in each of their corresponding fields, wherein the search module is configured to search each of the fields based on the recognized content and provide a plurality of listings as search results, wherein the search module is configured to analyze the recognized content to ascertain information therein other than a name in a desired listing, and wherein the search module is configured to discard one or more listings from the search results based on the information.

6. The computer system of claim 1 wherein each of the listings in the database comprise a sequence of character-based n-grams, and wherein the search module is configured to search the database based on the recognized content in the form of character-based n-grams.

7. The computer system of claim 1 wherein each of the listings in the database comprises phone n-grams, and wherein the search module is configured to search the database based on the recognized content in the form of phone-based n-grams.

8. The computer system of claim 1 wherein the search module is configured to search the database allowing inexact matches between terms in the recognized content and said at least one listing.

9. The computer system of claim 8 wherein the search module is configured to add expanded terms to the recognized content.

10. The computer system of claim 9 wherein the search module is configured to provide a plurality of listings as search results wherein listings attributed to the expanded terms are weighted in correspondence to acoustic similarity of the expanded terms to terms initially in the recognized content.

11. The computer system of claim 1 wherein one of the query recognizer or the search module is configured to modify the recognized content to ascertain a selected combination of terms from the recognized content.

12. The computer system of claim 1 wherein the search module is configured to provide a plurality of listings as search results;

each listing has an associated search score; and

wherein the search module is configured to combine the search scores with other scores and rerank the plurality of listings based on the combination of the search scores with the other scores.

13. The computer system of claim 1 wherein other scores are based on ranking provided by the query recognizer.

14. A computer-implemented method for searching listings in a database, the method comprising:

generating recognized content based on receipt of a user'"'"'s query; and

searching the database using a term frequency-inverse document frequency (Tf/Idf) algorithm based on the recognized content and providing at least one listing associated with the query.

15. The computer-implemented method of claim 14 and further comprising adding terms to the recognized content.

16. The computer-implemented method of claim 15 wherein adding terms includes eliminating duplicative terms by adding a unique term for one of the duplicative terms.

17. The computer-implemented method of claim 14 wherein the database is a structured database comprising a plurality of fields and wherein a plurality of listings have terms in each of their corresponding fields, and wherein searching comprises searching two or more fields using the recognized content.

18. The computer-implemented method of claim 14 wherein each of the listings in the database is n-gram based, and the method further comprising converting the recognized content into an n-gram based form corresponding to the form used in the database.

19. The computer-implemented method of claim 14 and wherein searching comprises obtaining a plurality of ranked listings as search results, the method further comprising reranking the plurality of listings of the search results based on other information.

20. The computer-implemented method of claim 14 wherein searching comprises obtaining a plurality of ranked listings as search results, the method further comprising adding terms to the recognized content, and reranking the listings where listings attributed to the added terms are weighted in correspondence to acoustic similarity of the added terms to terms initially in the recognized content.