MULTIDIMENSIONAL EXPERT BEHAVIOR EMULATION SYSTEM

• US 20090089241A1
• Filed: 12/09/2008
• Published: 04/02/2009
• Est. Priority Date: 10/26/2004
• Status: Abandoned Application

• Current Assignee: Applied Intelligence Solutions LLC (International Business Machines Corporation)
• Original Assignee: Applied Intelligence Solutions LLC (International Business Machines Corporation)
• Inventors: Kopiev, Ivan, Baytsur, Grigory, Zelinsky, Pavel, Belousov, Andrey, Grechko, Vitaly, Sidorkin, Oleg
• Application Number:
  US12/331,256
• Publication Number:
  US 20090089241A1
• Time in Patent Office: Days
• US Class Current: 706/60
• CPC Class Codes: G06N 5/047 Pattern matching networks; Rete networks

Abstract

An expert decision-making method is emulated based on a history of behaviors by experts in a variety of observed situations. The history of behaviors is built up from observations of actions taken by experts in analyzing a plurality of situations. Situation data representative of a situation to be processed is received, and situation features are extracted from the situation data. Each situation feature is associated with an expert behavior method used to process the situation. A behavior method is recognized from a pattern of situation features. Recognizing a behavior method is based on feature/method separation data in multidimensional space of features into so areas with each area associated with a method used by experts. Parameter values for parameters in the recognized behavior method are calculated based on the situation features. The calculation of parameter values is accomplished by recognizing parameter calculation rules and calculating the parameter values using the rules. A parameter calculation rule for each parameter in the behavior method is recognized from a pattern of situation features. Recognizing a parameter calculation rule is based on feature/parameter-calculation-rules separation data of multidimensional space of features into areas with each area associated with a parameter calculation rule used by experts. The recognized behavior method is executed on the situation data using the calculated parameter values to recommend a solution for the situation. The recommended solution has solution data representing a plan of action to provide the solution and remainder data representing unprocessed situation data. A test detects whether the remainder data is in a target range. If the remainder data is not in the target range, the actions to recommend a solution are repeated until the test detects the remainder data is in the target range.

Claims

1-10. -10. (canceled)

11. A computer method for emulating an expert decision-making process based on an history of behaviors by human experts in a variety of observed allocation situations to provide a user a recommended allocation solution to an allocation situation to be processed, the method comprising:

receiving situation data representative of an allocation situation to be processed;

extracting situation features from the situation data;

recognizing a behavior method from a pattern of situation features and identifying the recognized behavior method;

calculating parameter values based on the situation features for parameters in the recognized behavior method;

executing the recognized behavior method on the situation data using the parameter values to provide a recommended allocation solution for the allocation situation; and

displaying the recommended allocation solution to the user.

12. The method of claim 11 wherein the recommended allocation solution comprises solution data representing a plan of action to provide the allocation solution and remainder data representing unprocessed situation data.

13. The method of claim 12 further comprising:

detecting whether the remainder data is in a target range;

if the remainder data is not in the target range, recognizing another behavior method from a pattern of situation features from the remainder data and identifying the another recognized behavior method;

calculating parameter values, based on the situation features from the remainder data, for parameters in the another recognized behavior method; and

executing the another recognized behavior method on the remainder data using the parameter values in the another recognized behavior method to recommend an allocation solution for the allocation situation.

14. The method of claim 13 wherein the acts of recognizing, calculating and executing for another recognized method are repeated until the act of detecting detects the remainder data is in the target range.

15. The method of claim 11 wherein the displaying step further comprises:

displaying a strategy, the strategy being all the recognized behavior methods with parameter values and the solution data and remainder data for each recognized method.

16. The method of claim 11 wherein the act of extracting features bases the extraction on features that are significant in recognizing a behavior method.

17. The method of claim 11 wherein the act of recognizing a method is based on feature/method separation data in multidimensional space of features into areas with each area associated with a method used by experts.

18. The method of claim 11 wherein the act of calculating parameter values comprises:

recognizing a parameter calculation rule for each parameter in the recognized behavior method from a pattern of situation features; and

calculating the parameter values using parameter calculation rules using situation features.

19. The method of claim 18 wherein the act of recognizing a parameter calculation rule is based on feature/parameter-calculation-rules separation data in multidimensional space of features into areas with each area associated with a parameter calculation rule used by experts.

20. The method of claim 11 further comprising:

capturing observations of actions taken by experts in analyzing a plurality of allocation situations;

constructing behaviors from the observations, the behaviors indicating situation features and associated methods with parameter values.

21. The method of claim 20 further comprising:

learning a history of allocation situations and associated behavior methods from the behaviors; and

generating a profile of strategies from the history.

22. The method of claim 15 further comprising:

receiving adjustments to the strategy, the adjustments being at least one of changes in recognized methods and changes in parameter values.