 |
|
|
|
|
|
 Title: |
US5950189:
Retrieval system and method
[ Derwent Title ]
|
| Country: |
US United States of America
|
| |
| Inventor: |
Cohen, Edith; Berkeley Heights, NJ
Lewis, David Dolan; Summit, NJ
|
| Assignee: |
AT&T Corp, Middletown, NJ
other patents from AT&T CORP. (706518) (approx. 16,328)
News, Profiles, Stocks and More about this company
|
| Published / Filed: |
1999-09-07
/ 1997-01-02
|
| Application Number: |
US1997000775913
|
| IPC Code: |
Advanced:
G06F 17/30;
Core:
more...
IPC-7:
G06F 17/30;
|
| ECLA Code: |
G06F17/30T2P4V;
|
| U.S. Class: |
Current:
707/003;
707/004;
707/005;
707/E17.08;
Original:
707/003;
707/004;
707/005;
|
| Field of Search: |
707/001,2,3,4,5
|
| Priority Number: |
| 1997-01-02 |
US1997000775913 |
|
| Abstract: |
The invention is an improved retrieval system and method. Many pattern recognition tasks, including estimation, classification, and the finding of similar objects, make use of linear models. For example, many text retrieval systems represent queries as linear functions, and retrieve documents whose vector representation has a high dot product with the query. The fundamental operation in such tasks is the computation of the dot product between a query vector and a large database of instance vectors. Often instance vectors which have high dot products with the query are of interest. The invention relates to a random sampling based retrieval system that can identify, for any given query vector, those instance vectors which have large dot products, while avoiding explicit computation of all dot products.
|
| Primary / Asst. Examiners: |
Black, Thomas G.; Coby, Frantz
|
| INPADOC Legal Status: |
Show legal status actions
|
| Family: |
None
|
First Claim:
Show all 8 claims |
What is claimed is:
1. A retrieval system for retrieving data from a database, the database comprising records each having a set of attribute values and a record identifier, comprising:
- a database sampling unit, for sampling from the records in the database and generating a sampled representation of the database according to a probability distribution over the records of the database,
- the probability distribution being constructed as a function of the attribute values in the records, and
- the database sampling unit being constructed to generate a weighted, directed graph representing the database, the database sampling unit determining the probability distribution by random walks on the weighted, directed graph;
- a query input unit, for receiving a database query; and
- a query processing unit, operatively connected to the query input unit and to the database sampling unit, which applies the database query to the sampled representation of the database to return results.
|
| Background / Summary: |
Show background / summary
|
| Drawing Descriptions: |
Show drawing descriptions
|
| Description: |
Show description
|
| Forward References: |
Show 56 U.S. patent(s) that reference this one
|
|
