Searching a document using relevance feedback

• US 20080021891A1
• Filed: 07/19/2007
• Published: 01/24/2008
• Est. Priority Date: 07/19/2006
• Status: Active Grant

• Current Assignee: Ricoh Company Limited
• Original Assignee: Ricoh Company Limited
• Inventors: Ikeda, Tetsuya
• Granted Patent: US 7,769,771 B2
• Time in Patent Office: Days
• US Class Current: 707/3
• CPC Class Codes: G06F 16/3328 using graphical result space presentation or visualisation

Abstract

When a search request is input, an initial search result that matches the search request is displayed to a user. When a seed document set is selected by the user from the initial search result, an extension word set is extracted from the seed document set. Using the extension word set in addition to the search request, another search is performed.

Claims

1. A document search apparatus, comprising:

a search request input to input a search request including a keyword set;

a user interface to cause a display to display an initial search result to a user, the initial search result comprising a first retrieved document set retrieved from a plurality of documents stored in a document database using the keyword set of the search result;

an extension word extractor to extract an extension word set from a seed document set, the seed document set being selected by the user from the first retrieved document set of the initial search result; and

a database manager to search through the plurality of documents stored in the document database for a second retrieved document set using the keyword set of the search request and the extension word set.

2. The apparatus of claim 1, wherein a first retrieved document of the first retrieved document set is displayed as a symbol having a position on a coordinate system, the position being defined by a relevance value of the first retrieved document of the first retrieved document set, and at least one attribute value of the first retrieved document of the first retrieved document or a numerical value being converted from the at least one attribute value of the first retrieved document of the first retrieved document set.

3. The apparatus of claim 2, wherein the user interface is further operable to display, upon detecting a user instruction performed on the symbol being displayed, additional information regarding the first retrieved document represented by the symbol.

4. The apparatus of claim 2, wherein the user interface is operable to detect a user instruction for specifying at least one point in the coordinate system, and select the seed document set according to an area defined by the at least one point being specified by the user instruction.

5. The apparatus of claim 2, wherein the user interface is further operable to detect a user instruction for moving the position of the symbol to a desired position on the coordinate system, and store the desired position being specified by the user instruction in a corresponding manner with identification information of the first retrieved document represented by the symbol.

6. The apparatus of claim 2, wherein the extension word set is extracted based on a usefulness value of a word included in a seed document of the seed document set, the usefulness value being obtained using a weighting factor of the seed document of the seed document set.

7. The apparatus of claim 6, wherein the user interface is further operable to detect a user instruction for moving the position of the symbol to a desired position on the coordinate system before the first retrieved document represented by the symbol is selected as the seed document of the seed document set, and send information regarding the desired position to the extension word extractor together with identification information of the seed document, and wherein the extension word extractor is further configured to update the weighting factor of the seed document using the desired position.

8. The apparatus of claim 6, wherein the weighting factor of the seed document of the seed document set is previously assigned based on the relevance value of the seed document.

9. The apparatus of claim 2, wherein the user interface is further operable to detect a user instruction for changing a type of the at least one attribute value being displayed on the coordinate system, and change appearance of the initial search result according to the user instruction.

10. A method for searching a document, the method comprising:

inputting a search request including a keyword set;

searching through a plurality of documents stored in a document database for one or more first retrieved documents that match the keyword set of the search request to obtain an initial search result, the initial search result comprising a document identifier, a relevance value, and at least one attribute value for each one of the one or more first retrieved documents;

displaying the initial search result to a user;

selecting one or more seed documents as a seed document set from the one or more first retrieved documents according to a user instruction performed on the initial search result being displayed;

extracting an extension word set from the seed document set; and

searching through the plurality of documents stored in the document database for one or more second retrieved documents that match the keyword set of the search request and the extension word set.

11. The method of claim 10, wherein displaying the initial search result comprises:

converting the at least one attribute value to a numerical value when the at least one attribute value is not expressed in the form of numerical value;

defining a position of at least one of the one or more first retrieved documents on a coordinate system, using the relevance value and the at least one attribute value or the numerical value of the at least one of the one or more first retrieved documents; and

placing a symbol at the position being defined.

12. The method of claim 11, wherein displaying the initial search result further comprises:

storing the position being defined in a corresponding manner with the document identifier of the at least one of the one or more first retrieved documents.

13. The method of claim 11, wherein displaying the initial search result further comprises:

detecting a user instruction for requesting additional information regarding at least one of the one or more first retrieved documents, the user instruction being performed on the initial search result being displayed;

obtaining the additional information regarding the at least one of the one or more first retrieved documents; and

displaying the additional information to the user.

14. The method of claim 10, wherein selecting the one or more seed documents comprises:

detecting the user instruction, the user instruction specifying at least one point on the coordinate system; and

obtaining identification information of one or more symbols that are located within an area defined by the at least one point being specified by the user instruction,

wherein the one or more seed documents being selected as the seed document set are identified using the identification information being obtained.

15. The method of claim 10, wherein extracting the extension word set comprises:

segmenting at least one of the one or more seed documents into a plurality of words; and

obtaining a usefulness value of at least one of the plurality of words,

wherein the extension word set includes one or more words selected from the plurality of words based on the usefulness values.

16. The method of claim 15, wherein the usefulness value of at least one of the plurality of words is obtained using a weighting factor of the corresponding one of the one or more seed documents, the weighting factor being determined based on the relevance value of the corresponding one of the one or more seed documents.

17. The method of claim 11, wherein displaying the initial search result further comprises:

detecting a user instruction for moving the position of the symbol to a desired position on the coordinate system;

relocating the symbol at the desired position; and

storing the desired position in a corresponding manner with the document identifier of the at least one of the one or more first retrieved documents.

18. The method of claim 11, wherein displaying the initial search result further comprises:

detecting a user instruction for changing a type of the at least one attribute value of the first initial search result; and

changing appearance of the initial search result being displayed to the user according to the user instruction.

19. A computer readable medium storing computer instructions for performing a method for searching a document, the method comprising:

inputting a search request including a keyword set;

searching through a plurality of documents stored in a document database for one or more first retrieved documents that match the keyword set of the search request to obtain an initial search result, the initial search result comprising a document identifier, a relevance value, and at least one attribute value for each one of the one or more first retrieved documents;

displaying the initial search result to a user;

selecting one or more seed documents as a seed document set from the one or more first retrieved documents according to a user instruction performed on the initial search result being displayed;

extracting an extension word set from the seed document set; and

searching through the plurality of documents stored in the document database for one or more second retrieved documents that match the keyword set of the search request and the extension word set.

20. The medium of claim 19, wherein displaying the initial search result further comprises:

detecting a user instruction for moving the position of the symbol to a desired position on the coordinate system;

relocating the symbol at the desired position; and

storing the desired position in a corresponding manner with the document identifier of the at least one of the one or more first retrieved documents.