Query recognizer

• US 20050131872A1
• Filed: 12/16/2003
• Published: 06/16/2005
• Est. Priority Date: 12/16/2003
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Calbucci, Marcelo, Salas, Sally
• Application Number:
  US10/736,915
• Publication Number:
  US 20050131872A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3338 Query expansion

Abstract

A system and method for executing a query on a plurality of data sources which may include remotely located data sources having specialized data. These data sources might include, for example, data relating to retail sales, travel, entertainment and the like. The data source may also contain only data that is local to a given community or group of communities. The system includes a pre-processor which accepts an input query and modifies its form to provide a resultant query. This modification can include corrections in spelling for example. The system classifies the resultant query based on the presence of classifier terms and user provided classification data from previously submitted queries or recognized patterns in the query. The system then selects a subset of data sources upon which to execute the query based on this query classification. The query is then executed on the selected data sources.

Claims

1. A system for analyzing queries from a user and responding to said queries comprising:

a) a query processor for evaluating an input query and transmitting a modified form of said query to a data source for creating a response to the modified form of said query; and

b) a plurality of recognizers for evaluating the query or a modified form of the query and identifying data in the query;

c) said query processor including a recognizer broker for sending the query to a specified one or more of the plurality of recognizers.

2. The system of claim 1 wherein said recognizer broker categorizes the query as a certain type query with a certain degree of certainty.

3. The system of claim 1 wherein said query recognizer adds information to create a modified form of the input query based on data identified in said query.

4. The system of claim 3 wherein the query recognizer adds a category to the input query.

5. The system of claim 3 wherein the query recognizer adds information related to information contained in the input query.

6. The system of claim 1 additionally comprising a web crawler for searching documents and categorizing said documents and a database for providing results based upon an input from the query processor.

7. The system of claim 1 wherein the query processor modifies the query to include a category before transmitting the modified query to a data source.

8. The system of claim 1 wherein the query processor changes spelling of misspelled words in the query.

9. The system of claim 1 wherein one of the recognizers matches query input words with words stored in a database and categorizes said words with a confidence level.

10. The system of claim 9 wherein the database includes a history of user ratings for use in establishing the confidence level.

11. The system of claim 9 wherein the history of user ratings are based on user choices of previous results presented based on previously submitted queries.

12. The system of claim 10 wherein the database contains an index of words and an indication of a category of query for said words wherein multiple categories exist for the same word.

13. The system of claim 1 including a pattern recognizer for recognizing a pattern of characters in an input query for categorizing said query.

14. The system of claim 1 wherein one or more pattern recognizers contain a database of terms of specified languages for identification in said recognizer.

15. A system for analyzing queries from a user and responding to said queries with data comprising:

a) a query processor for evaluating the query and transmitting a form of said query to a data source for creating a response to the modified form of said query; and

b) a query recognizer for evaluating the query or a modified form of the query and identifying a query type based on tokens identified in said query;

c) a database that stores tokens and associates a confidence level for a type of query containing said tokens and provides a confidence level to said query recognizer when a match between a token in an input query and a token in said database is confirmed.

16. The system of claim 15 wherein the query processor modifies a form of the query before transmitting the modified form to said data source, said modified form including a category of query indicator.

17. A method of evaluating a query comprising:

a) evaluating an input query and transmitting a modified form of said query to a data source for creating a response to the modified form of said query by evaluating the query using a specified one or more of a plurality of query recognizers that evaluate tokens in the query to generate the modified form of the query.

18. The method of claim 17 wherein said input query is identified as a query of a certain type with a certain degree of certainty.

19. The method of claim 17 comprising adding information to create a modified form of the input query based on data identified in said input query.

20. The method of claim 17 comprising adding a category to the input query.

21. The method of claim 17 comprising adding information related to information contained in the input query.

22. The method of claim 17 additionally comprising searching for documents and categorizing said documents and maintaining a database for providing probabilities based upon an input derived from the query processor based on a content of the input query.

23. The method of claim 17 wherein the query processor changes spelling of misspelled words in the input query.

24. The method of claim 17 wherein one of the query recognizers matches query input words with words stored in a database and categorizes said words with a confidence level.

25. The method of claim 24 wherein the database includes a history of user ratings for use in establishing the confidence level.

26. The method of claim 25 wherein the history of user ratings are based on user choices of previous results presented based on previously submitted input queries.

27. The method of claim 26 wherein the database contains an index of words and an indication of a category of query for said words wherein multiple categories exist for the same word.

28. The method of claim 17 comprising recognizing a pattern of characters in an input query for modifying said query.

29. The method of claim 28 wherein one or more pattern recognizers contain a database of terms of specified languages for identification in said recognizer.

30. A computer readable medium having instructions for executing a query on a plurality of data sources comprising instructions for:

preprocessing the query to form a resultant query;

classifying the resultant query based on the presence of classifier terms and user provided classification data from previously submitted queries; and

selecting a subset of data sources upon which to execute the query based on query classification.