Topic net generation method and apparatus

• US 20050192959A1
• Filed: 05/04/2005
• Published: 09/01/2005
• Est. Priority Date: 01/23/2003
• Status: Active Grant

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Kitagawa, Eiji, Miyata, Ryosuke, Fukuoka, Toshiyuki
• Granted Patent: US 7,756,897 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: Y10S 707/944 Business related
• CPC Class Codes: Y10S 707/961 Associative

Abstract

A topic net management unit 1 reads one record portion of data from a collection of data in table format (S1). Subsequently, from among the field data included in the record that was read, the topic net management unit 1 generates, based on the definitions of the topic classes, nodes corresponding to the field data for which a corresponding topic class exists (S2). In addition, topics corresponding to those nodes are generated. If a node having the same topic is already generated, then the topic net management unit 1 does not newly generate a node having a duplicate topic (S2). The topic net management unit 1 mutually links the generated topics in accordance with the definitions between the topic classes (S3). Further, using a dialog information template, the topic net management unit 1 generates dialog information associated with the generated link (S4). The topic net management unit 1 performs Steps S1-S4 processing for all records included in the data in table format (S5). If there is a plurality of tables to generate the topic net, then the topic net management unit 1 performs Steps S1-S5 processing for all tables (S6).

Claims

1. A method of generating a topic net that includes nodes associated with topics and links that link each of the nodes, comprising the steps of:

(a) reading one record portion of data from a data table, wherein field data described in a plurality of fields is associated and stored;

(b) defining topic classes corresponding to all or a portion of said plurality of fields, and associations between each of the topic classes;

(c) generating the nodes with respect to field data among the data read in step (a), for which there exists a corresponding topic class, and topics corresponding to said nodes; and

(d) linking each of the nodes generated in step (c) in accordance with the associations between each of the topic classes defined in step (b).

2. The topic net generation method as recited in claim 1, wherein

step (c) includes stopping the generation of nodes having duplicate topics.

3. The topic net generation method as recited in claim 1, wherein

step (b) further defines a dialog information generating rule for generating dialog information that includes a message used to advance the dialog based on said topic net;

and

said dialog information generating rule generates said dialog information based on field data described in a field corresponding to two topic classes associated in step (b).

4. The topic net generation method as recited in claim 3, wherein

said dialog information generating rule further includes a direction that indicates from which node to which node, among the nodes of said two topic classes, the generated dialog information will be used to advance the dialog.

5. The topic net generation method as recited in claim 3, wherein

said dialog information generating rule further includes a message type that indicates what kind of property said message has.

6. A topic net generating apparatus that generates a topic net that includes nodes associated with topics and links that link each of the nodes, comprising:

a reading unit that reads one record portion of data from a data table, wherein field data described in a plurality of fields is associated and stored;

a defining unit that defines topic classes corresponding to all or a portion of said plurality of fields, and associations between each of the topic classes;

a node generating unit that generates the nodes with respect to field data among the data read by said reading unit, for which there exists a corresponding topic class, and topics corresponding to said nodes; and

a link generating unit that links each of the nodes generated by said node generating unit in accordance with the associations between each of the topic classes defined by said defining unit.

7. A topic net generating program that generates a topic net that includes nodes associated with topics and links that link each of the nodes, and that causes a computer to execute the steps of:

(a) reading one record portion of data from a data table, wherein field data described in a plurality of fields is associated and stored;

(b) defining topic classes corresponding to all or a portion of said plurality of fields, and associations between each of the topic classes;

(c) generating the nodes with respect to field data among the data read by said reading step, for which there exists a corresponding topic class, and topics corresponding to said nodes; and

(d) linking each of the nodes generated in step (c) in accordance with the associations between each of the topic classes defined in step (b).

8. A method of generating a topic net that includes nodes associated with topics and links that link each of the nodes, comprising the steps of:

(a) reading one tree portion of data from a data file wherein field data described in a plurality of fields and a tree structure having said fields are defined;

(b) defining topic classes corresponding to all or a portion of said plurality of fields, and associations between each of the topic classes;

(c) generating the nodes with respect to field data among the data read by said reading step, for which there exists a corresponding topic class, and topics corresponding to said nodes; and

(d) linking each of the nodes generated in step (c) in accordance with the associations between each of the topic classes defined in step (b).

9. The topic net generation method as recited in claim 8, wherein

step (c) includes stopping the generation of nodes having duplicate topics.

10. The topic net generation method as recited in claim 8, wherein

step (b) further defines a dialog information generating rule for generating dialog information that includes a message used to advance the dialog based on said topic net;

and

said dialog information generating rule generates said dialog information based on field data described in a field corresponding to two topic classes associated in step (b).

11. The topic net generation method as recited in claim 10, wherein

said dialog information generating rule further includes a direction that indicates from which node to which node, among the nodes of said two topic classes, the generated dialog information will be used to advance the dialog.

12. The topic net generation method as recited in claim 10, wherein

said dialog information generating rule further includes a message type that indicates what kind of property said message has.

13. A topic net generating apparatus that generates a topic net that includes nodes associated with topics and links that link each of the nodes, comprising:

a reading unit that reads one tree portion of data from a data file wherein field data described in a plurality of fields and a tree structure having said fields are defined;

a defining unit that defines topic classes corresponding to all or a portion of said plurality of fields, and associations between each of the topic classes;

a node generating unit that generates the nodes with respect to field data among the data read by said reading unit, for which there exists a corresponding topic class, and topics corresponding to said nodes; and

a link generating unit that links each of the nodes generated by said node generating unit in accordance with the associations between each of the topic classes defined by said defining unit.

14. A topic net generating program that generates a topic net that includes nodes associated with topics and links that link each of the nodes, and that causes a computer to execute the steps of:

(a) reading one tree portion of data from a data file wherein field data described in a plurality of fields and a tree structure having said fields are defined;

(b) defining topic classes corresponding to all or a portion of said plurality of fields, and associations between each of the topic classes;

(c) generating the nodes with respect to field data among the data read in step (a), for which there exists a corresponding topic class, and topics corresponding to said nodes; and

(d) linking each of the nodes generated in step (c) in accordance with the associations between each of the topic classes defined in step (b).

15. A topic net generation method that includes nodes associated with topics, and links that link each of the nodes, comprising the steps of:

(a) reading a data unit, a starting topic, and an ending topic that were correlated and stored;

(b) generating nodes corresponding to the read starting topic and ending topic, and topics corresponding to said nodes;

(c) generating a link from a node corresponding to said starting topic to a node corresponding to said ending topic; and

(d) generating dialog information, which includes a message used to advance the dialog based on said topic net and a message type that indicates the property of said message, based on said data unit, and associates said dialog information with the link generated in step (c);

wherein,

said data unit includes a text message, and a data type that indicates the property of said text message.

16. The topic net generation method as recited in claim 15, wherein

if the node corresponding to said starting topic and the node corresponding to said ending topic are identical nodes, then step (c) stops the generation of a link;

and

if the generation of a link is stopped, step (d) associates generated dialog information with said node.

17. The topic net generation method as recited in claim 16, wherein

if there exists a plurality of data units that include the same data type for identical links or identical nodes, then step (d) selects any one of the data units based on said supplementary information, and generates dialog information based on the selected data unit.

18. The topic net generation method as recited in claim 15, wherein

said data unit further includes supplementary information;

and

step (d) generates said dialog information based on said supplementary information.

19. The topic net generation method as recited in claim 15, wherein

step (b) stops the generation of nodes having duplicate topics.

20. A topic net generation apparatus that generates a topic net that includes nodes associated with topics, and links that link each of the nodes, comprising:

a reading unit that reads a data unit, a starting topic, and an ending topic that were correlated and stored;

a node generating unit that generates nodes corresponding to the read starting topic and ending topic, and topics corresponding to said nodes;

a link generating unit that generates a link from a node corresponding to said starting topic to a node corresponding to said ending topic; and

a dialog information generating unit that generates dialog information, which includes a message used to advance the dialog based on said topic net and a message type that indicates the property of said message, based on said data unit, and associates said dialog information with the link generated by said link generating unit;

wherein,

said data unit includes a text message, and a data type that indicates the property of said text message.

21. A topic net generation program that generates a topic net that includes nodes associated with topics, and links that link each of the nodes, and that causes a computer to execute the steps of:

(a) correlating and reading a data unit, which includes a text message and a message type that indicates the property of said text message, a starting topic, and an ending topic;

(b) generating nodes corresponding to the read starting topic and ending topic, and topics corresponding to said nodes;

(c) generating a link from a node corresponding to said starting topic to a node corresponding to said ending topic; and

(d) generating dialog information, which includes a message used to advance the dialog based on said topic net and a message type that indicates the property of said message, based on said data unit, and associates said dialog information with the link generated in step (c).