System and method for case-based reasoning

• US 7,225,176 B2
• Filed: 03/26/2003
• Issued: 05/29/2007
• Est. Priority Date: 03/26/2003
• Status: Active Grant

• Current Assignee: CaseBank Technologies Inc.
• Original Assignee: CaseBank Technologies Inc.
• Inventors: Atamer, Allen
• Primary Examiner(s): Hirl; Joseph P
• Application Number:
  US10/396,388
• Publication Number:
  US 20040193560A1
• Time in Patent Office: 1,525 Days
• Field of Search: 706/46, 706/50, 706/45, 707/102
• US Class Current: 706/46
• CPC Class Codes: G06F 16/24578 using ranking

Abstract

A case based reasoning system and method for matching a problem case to at least one solved case. A solved case database stores solved case data correlated to a plurality of solved cases. An attribute database stores a set of attributes and observation cost data corresponding to each of the attributes, such that the observation cost data corresponding to an attribute is correlated to a cost of determining an attribute value corresponding to the attribute. A set of potential solved cases is determined. A set of relevant attributes for which at least one potential solved case has an attribute value and for which no corresponding problem attribute value has been input is determined and ranked based on the observation cost data for each relevant attribute.

Claims

1. A method for matching at least one solved case to a problem case, the method comprising:

(a) storing solved case data correlated to a plurality of solved cases, wherein the solved case data comprises a set of attributes corresponding to each of the solved cases;

(b) storing attribute data corresponding to a set of attributes;

(c) storing observation cost data corresponding to each attribute in the set of attributes, wherein the observation cost data corresponding to an attribute is correlated to a cost of determining an attribute value corresponding to the attribute;

(d) receiving problem attribute values for attributes correlated to the problem case;

(e) determining a list of potential solved cases from said solved case data by comparing the problem attribute values to the attribute values of the plurality of solved cases;

(f) determining a list of relevant attributes for which at least one potential solved case has an attribute value and for which no corresponding problem attribute value has been input; and

(g) ranking said list of relevant attributes based at least in part on the observation cost data.

2. The method of claim 1, further comprising the step of:

(h) ranking said list of potential solved cases.

3. The method of claim 1, further comprising the step of:

(i) receiving a problem attribute value corresponding to a relevant attribute.

4. The method of claim 3, wherein step (i) comprises determining the problem attribute value.

5. The method of claim 3, further comprising repeating steps (f) and (g) following step (i).

6. The method of claim 2, further comprising the step of:

(i) receiving a problem attribute value corresponding to a relevant attribute;

and wherein steps (f) through (h) are repeated following step (i).

7. The method of claim 1, wherein the step of ranking said list of relevant attributes comprises determining a ranking value for each relevant attribute such that said ranking value is correlated to the observation cost data for that relevant attribute.

8. The method of claim 1, wherein step (a) further comprises storing case frequency data corresponding to each solved case, wherein said case frequency data corresponding to a solved case is correlated to the frequency of the solved case occurring relative to the frequency of other solved cases occurring.

9. A case-based reasoning system for matching at least one solved case to a problem case, wherein the reasoning system comprises:

(a) a solved case database comprising solved case data correlated to a plurality of solved cases, wherein the solved case data comprises a set of attribute values corresponding to each of the solved cases;

(b) an attribute database comprising a set of attributes and observation cost data corresponding to each of the attributes, wherein the observation cost data corresponding to an attribute is correlated to a cost of determining an attribute value corresponding to the attribute;

(c) an input device for inputting problem attribute values correlated to the problem case;

(d) a processor programmed to;

(i) determine a set of at least one potential solved case from said solved case database,

(ii) determine a set of relevant attributes for which at least one potential solved case has an attribute value and for which no corresponding problem attribute value has been input,

(iii) determine a ranking value for each relevant attribute such that said ranking value is correlated to the observation cost data for that relevant attribute, and

(iv) rank said set of relevant attributes; and

(e) an output device for displaying the list of potential solved cases and for displaying the set of ranked relevant attributes.

10. The case-based reasoning system as claimed in claim 9, wherein the processor is programmed to determine the set of potential solved cases by comparing the problem attribute values to the attribute values of the plurality of solved cases.

11. The case-based reasoning system as claimed in claim 10, wherein the processor is further programmed to rank the set of potential solved cases based at least in part on the similarity of the problem attribute values to the attribute values of each potential solved case.

12. The case-based reasoning system as claimed in claim 9, wherein the solved case database further comprises case frequency data corresponding to each solved case, wherein said case frequency data corresponding to a solved case is correlated to the frequency of the solved case occurring relative to the frequency of other solved cases occurring, and wherein the processor is further programmed to rank the potential set of solved cases based at least in part on the case frequency data.

13. A method for matching at least one solved case to a problem case in a case-based reasoning system, wherein the reasoning system comprises:

(a) a solved case database comprising solved case data correlated to a plurality of solved cases, wherein the solved case data comprises a set of attribute values corresponding to each of the solved cases;

(b) an attribute database comprising a set of attributes and observation costs data corresponding to each of the attributes, wherein the observation cost data corresponding to an attribute is correlated to a cost of determining an attribute value corresponding to the attribute;

wherein the method comprises the steps of;

(i) receiving problem attribute values correlated to the problem case;

(ii) determining a list of potential solved cases from said solved case database;

(iii) determining a list of relevant attributes for which at least one potential solved case has an attribute value and for which no corresponding problem attribute value has been input; and

(iv) determining a ranking value for each relevant attribute such that said ranking value is correlated to the observation cost data for that relevant attribute.

14. A method of creating data for use in a case-based reasoning system, the method comprising:

(a) storing on computer-readable media solved case data correlated to a plurality of solved cases, wherein the solved case data comprises a set of attribute values corresponding to each of the solved cases;

(b) storing on computer-readable media attribute data corresponding to a set of attributes;

(c) receiving observation cost data corresponding to each of the attributes in the set, wherein the observation cost data corresponding to an attribute is correlated to a cost of determining an attribute value corresponding to the attribute; and

(d) storing on computer-readable media the observation cost data.

15. The method of claim 14, wherein step (a) further comprises storing on computer-readable media case frequency data corresponding to each solved case, wherein said case frequency data corresponding to a solved case is correlated to the frequency of the solved case occurring relative to the frequency of other solved cases occurring.