Document detection system using detection result presentation for facilitating user's comprehension

• US 5,649,193 A
• Filed: 11/17/1995
• Issued: 07/15/1997
• Est. Priority Date: 03/12/1993
• Status: Expired due to Term

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Mizutani, Yumi, Hirakawa, Hideki, Sumita, Kazuo, Miike, Seiji
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): LEWIS, CHERYL RENEA
• Application Number:
  US08/560,512
• Time in Patent Office: 606 Days
• Field of Search: 395/600, 395/575, 395/425
• US Class Current: 1/1
• CPC Class Codes: G06F 16/26 Visual data mining; Browsing structured data
• CPC Class Codes: G06F 16/3332 Query translation
• CPC Class Codes: G06F 16/338 Presentation of query results
• CPC Class Codes: G06F 16/34 Browsing; Visualisation therefor
• CPC Class Codes: G06F 16/93 Document management systems
• CPC Class Codes: Y10S 707/99942 Manipulating data structure, e.g. compression, compaction, compilation
• CPC Class Codes: Y10S 707/99944 Object-oriented database structure
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing

Abstract

A document detection system using a detection result presentation that can facilitate a user'"'"'s quick comprehension of the relevancy of each detected document. In this system, a user'"'"'s input for commanding a document detection is analyzed to extract keywords and viewpoints relevant to each keyword contained in the user'"'"'s input, and a detection command is constructed from the keywords and the viewpoints extracted from the user'"'"'s input. Then, the detection operation is executed to detect those documents among a plurality of stored documents which match with the constructed detection command as detected documents for a detection result. The detection result can be presented in the multi-dimensional display formed by setting the viewpoints to axes with the detection command as an origin and using distances of the detected documents with respect to the origin for each viewpoint as coordinates of the detected documents with respect to each axis representing each viewpoint.

Claims

1. A document detection system, comprising:

document memory means for storing a plurality of documents;

input means for entering user'"'"'s input, the user'"'"'s input requesting a document detection to be carried out in the documents stored in the document memory means;

input analysis means for analyzing the user'"'"'s input entered by the input means to extract keywords and viewpoints from the user'"'"'s input, each viewpoint indicating a context in which a keyword is used in the user'"'"'s input, and for constructing a detection command from the keywords and the viewpoints extracted from the user'"'"'s input; and

detection means for detecting those documents stored in the document memory means which match with the detection command constructed by the input analysis means as detected documents of a detection result.

2. The system of claim 1, wherein the input analysis means extracts viewpoints relevant to a particular keyword according to prescribed viewpoint extraction rules specifying expression patterns by matching a part of the user'"'"'s input containing the particular keyword with the expression patterns and extracting a viewpoint corresponding to the expression patterns as a viewpoint relevant to the particular keyword.

3. The system of claim 1, wherein the detection means detects the detected documents by searching a set of each keyword and a viewpoint relevant to said each keyword through a keyword index enlisting all keywords of the documents stored in the document memory means in relation to the viewpoints.

4. A document detection system, comprising:

document memory means for storing a plurality of documents;

input means for entering user'"'"'s input, the user'"'"'s input requesting a document detection to be carried out in the documents stored in the document memory means;

input analysis means for analyzing the user'"'"'s input entered by the input means to extract keywords and viewpoints from the user'"'"'s input, each viewpoint indicating a context in which a keyword is used in the user'"'"'s input, and for constructing a detection command from the keywords and the viewpoints extracted from the user'"'"'s input;

detection means for detecting those documents stored in the document memory means which match with the detection command constructed by the input analysis means as detected documents of a detection result;

distance calculation means for calculating distances of the detected documents detected by the detection means with respect to the detection command, for each viewpoint; and

detection result display means for presenting a detection result display indicating the detection result obtained by the detection means, in a multi-dimensional display formed by setting different viewpoints to respective axes with the detection command as an origin and using the distances of the detected documents for said each viewpoint calculated by the distance calculation means as coordinates of the detected documents with respect to each axis representing said each viewpoint.

5. The system of claim 4, further comprising:

record management means for managing detection records each comprised of a detection command and a corresponding detection result; and

detection record display means for presenting a detection record display indicating relations among the detection records managed by the detection record management means in a form of a tree structure in which each detection record is represented by a node in the tree structure.

6. The system of claim 5, wherein the detection record display means commands the detection result display means to present a detection result display for a detection record represented by a node specified by a user on the detection record display.

7. The system of claim 4, further comprising:

means for entering accept/reject data for indicating an acceptance/rejection of each detected document in the detection result display, wherein the detection result display means indicates those detection documents which are accepted by the accept/reject data distinguishably from those detection documents which are rejected by the accept/reject data.

8. The system of claim 7, wherein the detection means constructs a new detection command according to the accept/reject data, and re-detects those documents stored in the document memory means which match the new detection command.

9. The system of claim 4, further comprising:

browsing means for displaying data indicative of each detected document specified by a user on the detection result display, wherein the detection result display means indicates each detected document for which the browsing means displays the data distinguishably from other detected documents on the detection result display.

10. The system of claim 9, wherein the browsing means displays at least one of a summary, a keyword list, and a viewpoint list as the data indicative of each detected document.

11. The system of claim 9, wherein the browsing means generates a window for separately displaying the data indicative of each detected document specified by the user on the detection result display, and arranges a plurality of windows for displaying the data for a plurality of the detected documents specified by the user in a desired order specified by the user.

12. The system of claim 4, wherein the detection result display means rotates the detection result display according to an input for specifying a rotation entered by a user on the detection result display.

13. The system of claim 4, wherein the detection result display means enlarges/contracts the detection result display according to an input for specifying a range in the multi-dimensional display entered by a user on the detection result display.

14. The system of claim 4, wherein the detection result display means presents the detection result display in a cluster display mode in which a multi-dimensional space of the multi-dimensional display is divided into a number of sub-spaces, and the detected documents located within each sub-space is represented collectively as a cluster.

15. The system of claim 14, wherein the detection result display means also presents the detection result display in a document display mode which includes the multi-dimensional display for only a cluster specified by a user on the detection result display in the cluster display mode.

16. The system of claim 14, wherein the detection result display means changes the detection result display in the cluster display mode according to a number of the detected documents contained in each cluster.

17. The system of claim 14, wherein the detection result display means enters accept/reject data for indicating an acceptance/rejection of all the detected documents in each cluster in the detection result display in the cluster display mode collectively.

18. A method of document detection, comprising the steps of:

analyzing a user'"'"'s input for commanding a document detection to extract keywords and viewpoints relevant to each keyword from the user'"'"'s input, the viewpoints each indicating a context in which said each key word is used in the user'"'"'s input;

constructing a detection command from the keywords and the viewpoints extracted from the user'"'"'s input at the analyzing step; and

detecting those documents among a plurality of stored documents which match with the detection command constructed at the constructing step as detected documents of a detection result.

19. A method of document detection, comprising the steps of:

analyzing a user'"'"'s input for commanding a document detection to extract keywords and viewpoints relevant to each keyword from the user'"'"'s input, the viewpoints each indicating a context in which said each keyword is used in the user'"'"'s input;

constructing a detection command from the keywords and the viewpoints extracted from the user'"'"'s input at the analyzing step;

detecting those documents among a plurality of stored documents which match with the detection command constructed at the constructing step as detected documents of a detection result;

calculating distances of the detected documents obtained at the executing step with respect to the detection command for each viewpoint; and

presenting a detection result display indicating the detection result obtained at the detecting step, in a multi-dimensional display formed by setting the viewpoints to respective axes with the detection command as an origin and using the distances of the detected commands for each viewpoint calculated at the calculating step as coordinates of the detected documents with respect to each axis representing each viewpoint.

20. The method of claim 19, wherein the presenting step presents the detection result display in a cluster display mode in which a multi-dimensional space of the multi-dimensional display is divided into a number of sub-spaces, and the detected documents located within each sub-space is represented collectively as a cluster.

21. A document detection system, comprising:

a memory storing a plurality of documents;

an input device for receiving a user command, the user command requesting a document detection in accordance with a user request;

a processing unit connected to receive the user command from the input device, wherein the processing unit

extracts one or more keywords from the user request and for each keyword determines one or more corresponding viewpoints indicating a context in which the keyword is used in the user request,

generates a detection command based on the extracted keywords and the corresponding viewpoints, and

identifies documents among the plurality of documents stored in the memory which satisfy the detection command as detected documents; and

a display device connected to the processing unit to display a representation of the detected documents as a function of the viewpoints used to generate the detection command.