Information filtering apparatus and method for retrieving a selected article from information sources

• US 6,052,714 A
• Filed: 12/13/1996
• Issued: 04/18/2000
• Est. Priority Date: 12/14/1995
• Status: Expired due to Fees

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Kajiura, Masahiro, Miike, Seiji, Ono, Kenji, Sumita, Kazuo, Sakai, Tetsuya
• Primary Examiner(s): Matar, Ahmad F.
• Assistant Examiner(s): Barot, Bharat
• Application Number:
  US08/766,859
• Time in Patent Office: 1,222 Days
• Field of Search: 395/898, 395/500, 395/200.3, 395/200.31, 395/200.32, 395/200.35, 395/200.36, 395/200.47, 395/182.01, 395/200, 455/4.2, 455/5.1, 348/2, 348/7-12, 709/200-203, 709/204-207, 709/217-219, 709/226, 707/1-6, 707/10, 707/100-104, 713/200, 713/201
• US Class Current: 709/217
• CPC Class Codes: G06Q 30/06 Buying, selling or leasing transactions
• CPC Class Codes: Y10S 707/99931 Database or file accessing

Abstract

An information filtering apparatus for providing articles to a user includes an article extraction section that retrieves articles from articles stored in an article storage section using a user-profile. An excerpt generating section generates excerpts of the retrieved article by calculating similarities between the retrieved articles and the user-profile to generate the excerpt including the user'"'"'s necessary information. An abstract generating section combines excerpts to form an abstract. The user-profile for each user in the user-profile storage section is modified in accordance with user feedback.

Claims

1. An information filtering apparatus for providing articles to a user, comprising:

retrieval means for retrieving a selected article from among articles, each having at least one paragraph, received from information sources according to retrieval conditions of the user, the retrieval conditions of the user including a profile in which words included in articles required for each user and a weight of the words are described, the retrieval means further including

means for calculating similarities between each paragraph of the articles received from the information sources and the words in the profile, and

means for extracting a paragraph which has a largest similarity among similarities of all paragraphs;

generation means for generating an abstract of the selected article retrieved by said retrieving means according to said retrieval conditions, the abstract including the extracted paragraph; and

sending means for sending the abstract generated by said generating means to the user.

2. An information filtering apparatus according to claim 1, further comprising means for extracting a plurality of paragraphs in order of similarity with the profile.

3. An information filtering apparatus according to claim 2, wherein the number of extracted paragraphs equals the total number of paragraphs divided by any integer.

4. An information filtering apparatus according to claim 1, further comprising storage means for storing the articles received from the information sources.

5. An information filtering apparatus according to claim 4, wherein said articles stored in the storage means each include an article ID, a publisher, a headline, and text.

6. An information filtering apparatus according to claim 1, wherein said generation means includes means for extracting a first paragraph of the article retrieved by said retrieval means.

7. An information filtering apparatus for providing an extract of an article to a user, comprising;

means for storing a user profile;

means for selecting an article from a plurality of articles in accordance with the user profile;

means for extracting a portion of the selected article;

means for sending the portion to the user; and

means for receiving an evaluation of the portion from the user and for modifying the user profile in accordance with the evaluation.

8. An information filtering apparatus according to claim 7, wherein said evaluation is represented that an interested portion is marked "o", an uninterested portion is marked "x", and the portion either interested or not is non-marked.

9. An information filtering apparatus for providing articles to a user, comprising:

retrieval means for retrieving a selected article from among articles received from information sources according to retrieval conditions of each user, the retrieval conditions of the user including a profile in which words included in articles required for the user and a weight of the words are described;

a plurality of generation means each for generating an excerpt of the article retrieved by said retrieval means according to said retrieval conditions and features of articles;

feature extraction means for extracting a feature of the selected article retrieved by said retrieval means;

selection means for selecting one of the generating means from said plurality of generation means according to the feature of the article extracted by the feature extraction means; and

sending means for sending the excerpt of the article generated by the selected generation means to the user.

10. An information filtering apparatus according to claim 9, wherein the feature of the article is the number of characters of the text of the article.

11. An information filtering apparatus according to claim 9, wherein the feature of the article is the number of paragraphs of the text of the article.

12. An information filtering apparatus according to claim 9, wherein said selection means selects of the generation means by comparing the feature of the article with the feature of an excerpt generated by each said generation means.

13. An information filtering apparatus according to claim 9, further comprising means for calculating a similarity between each paragraph of the article retrieved by said retrieval means and words of said retrieval condition;

and

means for extracting a paragraph which has the largest similarity among similarities of all paragraphs.

14. An information filtering apparatus according to claim 13, further comprising means for extracting a plurality of paragraphs in order of similarity.

15. An information filtering apparatus according to claim 14, wherein the number of extracted paragraphs is the total number of the all paragraphs divided by any integer.

16. An information filtering apparatus according to claim 9, further comprising means for receiving an evaluation for the excerpt of the articles from the user;

and

means for modifying the retrieval conditions according to the evaluation.

17. An information filtering apparatus according to claim 16, wherein said evaluation is represented that an interested article is marked "o", an uninterested article is marked "x", and the article either interested or not is non-marked.

18. An information filtering method for providing articles to a user, comprising the steps of:

retrieving a selected article from among articles from information sources according to retrieval conditions for the user, the retrieval conditions of the user including a profile in which words included in articles required for each user and a weight of the words are described, the retrieval step further including the sub-steps of

calculating similarities between each paragraph of the articles received from the information sources and the words in the profile, and

extracting a paragraph which has a largest similarity among similarities of all paragraphs;

generating an abstract of the selected article retrieved according to said retrieval conditions, the abstract including the extracted paragraph; and

sending the abstract to the user.

19. An information filtering method for providing articles to a user, comprising the steps of:

retrieving a selected article from among articles from information sources according to retrieval conditions for the user, the retrieval conditions of the user including a profile in which words included in articles required for the user and a weight of the words are described;

extracting a feature of the selected article;

generating an abstract of the selected article according to said retrieval conditions and said feature of the article extracted in the extracting step, the abstract of the selected article including a paragraph of the selected article that is most similar to the words in the user profile; and

sending the abstract to the user.

20. A computer readable memory containing computer-readable instructions, comprising:

instruction means for causing a computer to retrieve a selected article from among articles received from information sources according to retrieval conditions for a user, the retrieval conditions of the user including a profile in which words included in articles required for each user and a weight of the words are described, the retrieval step further including the sub-steps of

calculating similarities between each paragraph of the articles received from the information sources and the words in the profile, and

extracting a paragraph which has a largest similarity among similarities of all paragraphs;

instruction means for causing a computer to generate an abstract of the selected article according to the retrieval condition, the abstract including the extracted paragraph; and

instruction means for causing a computer to send the abstract to the user.

21. A computer readable memory containing computer-readable instructions, comprising:

instruction means for causing a computer to retrieve a selected article from among articles from information sources according to retrieval conditions for a user, the retrieval conditions of the user including a profile in which words included in articles required for the user and a weight of the words are described;

instruction means for causing a computer to extract a feature of the selected article;

instruction means for causing a computer to generate an abstract of the selected article according to the retrieval conditions and the feature of the selected article, the abstract of the selected article including a paragraph of the selected article that is most similar to the words in the user profile; and

instruction means for causing a computer to send the abstract to the user.