Scheme for filtering documents on network using relevant and non-relevant profiles

• US 7,152,063 B2
• Filed: 02/28/2001
• Issued: 12/19/2006
• Est. Priority Date: 03/13/2000
• Status: Expired due to Fees

• Current Assignee: DDI Corporation (TTM Technologies Incorporated)
• Original Assignee: DDI Corporation (TTM Technologies Incorporated)
• Inventors: Matsumoto, Kazunori, Inoue, Naomi, Hashimoto, Kazuo, Hoashi, Keiichirou
• Primary Examiner(s): Ali, Mohammad
• Assistant Examiner(s): Pannala, Sathyanarayan
• Application Number:
  US09/795,841
• Publication Number:
  US 20010032204A1
• Time in Patent Office: 2,120 Days
• Field of Search: 707 1- 10, 707 5- 7
• US Class Current: 1/1
• CPC Class Codes: G06F 16/335 Filtering based on additional data, e.g. user or group profiles filtering in web context G06F16/9535, G06F16/9536
• CPC Class Codes: Y10S 707/99931 Database or file accessing
• 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 707/99936 Pattern matching access
• CPC Class Codes: Y10S 707/99937 Sorting
• CPC Class Codes: Y10S 707/99938 Concurrency, e.g. lock management in shared database
• CPC Class Codes: Y10S 707/99939 Privileged access

Abstract

A document filtering scheme capable of reducing the number of erroneously selected non-relevant documents without reducing the number of correctly selected relevant documents is disclosed. a non-relevant profile expressing non-relevant documents that are judged as having high similarities with respect to a relevant profile expressing a user'"'"'s request but not matching with the user'"'"'s request is utilized in addition to the relevant profile, such that each search target document for which the similarity with respect to the relevant profile is higher than a prescribed threshold for the relevant profile and the similarity with respect to the non-relevant profile is higher than a prescribed threshold for the non-relevant profile is removed as a non-relevant document.

Claims

1. A document filtering method using a computer for extracting documents matching with a relevant profile expressing a user'"'"'s request in terms of features of relevant documents that were correctly selected in the past by the computer from search target documents and presenting extracted documents to a user through the computer, comprising the steps of:

producing a non-relevant profile expressing non-relevant documents in terms of features of documents that were judged in the past as having high similarities with respect to the relevant profile by the computer but not matching with the user'"'"'s request;

calculating a first similarity between the relevant profile and each search target document, and comparing the first similarity with a prescribed first threshold for the relevant profile;

calculating a second similarity between the non-relevant profile and each search target document for which the first similarity is higher than the prescribed first threshold, and comparing the second similarity with a prescribed second threshold for the non-relevant profile;

removing each search target document for which the first similarity is higher than the prescribed first threshold and the second similarity is higher than the prescribed second threshold as a non-relevant document, and selecting each search target document for which the first similarity is higher than the prescribed first threshold and the second similarity is not higher than the prescribed second threshold as a document to be presented to the user; and

updating the relevant profile and the non-relevant profile according to a feedback from the user regarding relevance of documents presented to the user.

2. The document filtering method of claim 1, wherein the updating step updates the non-relevant profile by:

extracting words expressing features of a selected document according to word contributions, from words appearing in a selected document;

calculating a contribution of each extracted word;

calculating a score of each extracted word by multiplying a weight to the word contribution of each extracted word; and

an updating the non-relevant profile by subtracting the score of each extracted word from the non-relevant profile when the selected document is a relevant document, or by adding the score of each extracted word to the non-relevant profile when the selected document is a non-relevant document.

3. The document filtering method of claim 1, wherein the updating step updates the non-relevant profile also according to a pseudo feedback regarding those search target documents that are regarded as non-relevant documents.

4. The document filtering method of claim 1, wherein the removing step also removes each search target document for which the first similarity is not higher than the prescribed first threshold as a non-relevant document.

5. A document filtering system using a computer for extracting documents matching with a relevant profile expressing a user'"'"'s request in terms of features of relevant documents that were correctly selected in the past by the computer from search target documents and presenting extracted documents to a user through the computer, comprising:

a non-relevant profile production unit configured to produce a non-relevant profile expressing non-relevant documents in terms of features of documents that were judged in the past as having high similarities with respect to the relevant profile by the computer but not matching with the user'"'"'s request;

a relevant profile similarity calculation unit configured to calculate a first similarity between the relevant profile and each search target document, and compare the first similarity with a prescribed first threshold for the relevant profile;

a non-relevant profile similarity calculation unit configured to calculate a second similarity between the non-relevant profile and each search target document for which the first similarity is higher than the prescribed first threshold, and compare the second similarity with a prescribed second threshold for the non-relevant profile;

a relevance judgment unit configured to remove each search target document for which the first similarity is higher than the prescribed first threshold and the second similarity is higher than the prescribed second threshold as a non-relevant document, and select each search target document for which the first similarity is higher than the prescribed first threshold and the second similarity is not higher than the prescribed second threshold as a document to be presented to the user; and

a profile updating unit configured to update the relevant profile and the non-relevant profile according to a feedback from the user regarding relevance of documents presented to the user.

6. The document filtering system of claim 5, wherein the profile updating unit is configured to update the non-relevant profile by:

extracting words expressing features of a selected document according to word contributions, from words appearing in a selected document;

calculating a contribution of each extracted word;

calculating a score of each extracted word by multiplying a weight to a word contribution of each extracted word; and

an updating the non-relevant profile by subtracting the score of each extracted word from the non-relevant profile when the selected document is a relevant document, or by adding the score of each extracted word to the non-relevant profile when the selected document is a non-relevant document.

7. The document filtering system of claim 5, wherein the profile updating unit is configured to update the non-relevant profile also according to a pseudo feedback regarding those search target documents that are regarded as non-relevant documents.

8. The document filtering system of claim 5, wherein the relevance judgment unit is configured to also remove each search target document for which the first similarity is not higher than the prescribed first threshold as a non-relevant document.

9. A computer usable medium having computer readable program codes embodied therein for causing a computer to function as a document filtering system for extracting documents matching with a relevant profile expressing a user'"'"'s request in terms of features of relevant documents that were correctly selected in the past by the computer from search target documents and presenting extracted documents to a user through the computer, the computer readable program codes include:

a first computer readable program code for causing said computer to produce a non-relevant profile expressing non-relevant documents in terms of features of documents that were judged in the past as having high similarities with respect to the relevant profile by the computer but not matching with the user'"'"'s request;

a second computer readable program code for causing said computer to calculate a first similarity between the relevant profile and each search target document, and compare the first similarity with a prescribed first threshold for the relevant profile;

a third computer readable program code for causing said computer to calculate a second similarity between the non-relevant profile and each search target document for which the first similarity is higher than the prescribed first threshold, and compare the second similarity with a prescribed second threshold for the non-relevant profile;

a fourth computer readable program code for causing said computer to remove each search target document for which the first similarity is higher than the prescribed first threshold and the second similarity is higher than the prescribed second threshold as a non-relevant document, and select each search target document for which the first similarity is higher than the prescribed first threshold and the second similarity is not higher than the prescribed second threshold as a document to be presented to the user; and

a fifth computer readable program code for causing said computer to update the relevant profile and the non-relevant profile according to a feedback from the user regarding relevance of documents presented to the user.

10. The computer usable medium of claim 9, wherein the fifth computer readable program code causes said computer to update the non-relevant profile by:

extracting words expressing features of a selected document according to word contributions, from words appearing in a selected document;

calculating a contribution of each extracted word;

calculating a score of each extracted word by multiplying a weight to the word contribution of each extracted word; and

an updating the non-relevant profile by subtracting the score of each extracted word from the non-relevant profile when the selected document is a relevant document, or by adding the score of each extracted word to the non-relevant profile when the selected document is a non-relevant document.

11. The computer usable medium of claim 9, wherein the fifth computer readable program code causes said computer to update the non-relevant profile also according to a pseudo feedback regarding those search target documents that are regarded as non-relevant documents.

12. The computer usable medium of claim 9, wherein the fourth computer readable program code causes said computer to also remove each search target document for which the first similarity is not higher than the prescribed first threshold as a non-relevant document.