INCREASING SIGNAL TO NOISE RATIO FOR CREATION OF GENERALIZED AND ROBUST PREDICTION MODELS

• US 20130268469A1
• Filed: 04/03/2013
• Published: 10/10/2013
• Est. Priority Date: 04/06/2012
• Status: Active Grant

• Current Assignee: Applied Materials Incorporated
• Original Assignee: Applied Materials Incorporated
• Inventors: Moyne, James, Sharma, Deepak, Armer, Helen R.
• Granted Patent: US 9,275,334 B2
• Time in Patent Office: Days
• US Class Current: 706/12
• CPC Class Codes: G05B 19/00 Programme-control systems
• CPC Class Codes: G06F 30/367 Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 20/20 Ensemble learning
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition

Abstract

A computer system iteratively executes a decision tree-based prediction model using a set of input variables. The iterations create corresponding rankings of the input variables. The computer system generates overall variables contribution data using the rankings of the input variables and identifies key input variables based on the overall variables contribution data.

Claims

1. A method comprising:

executing a decision tree-based prediction model for a plurality of iterations using a plurality of input variables, wherein the plurality of iterations create corresponding rankings of the plurality of input variables;

generating overall variables contribution data using the rankings of the plurality of input variables; and

identifying a plurality of key input variables based on the overall variables contribution data.

2. The method of claim 1, wherein generating the variables contribution data is based on at least one of a rules ensemble algorithm, a support vector machine algorithm, linear regression algorithm, or partial least squares algorithm.

3. The method of claim 1, wherein executing the decision tree-based prediction model comprises:

assigning a different weight to at least one of the plurality of input variables for the plurality of iterations.

4. The method of claim 3, wherein the different weight is at least one of assigned using on at least one of an order of the plurality of input variable in a vector or assigned randomly.

5. The method of claim 1, wherein generating the overall variables contribution data comprises:

generating a plurality of variables contribution results corresponding to the plurality of iterations, the plurality of variables contribution results indicating a plurality of contributions corresponding to the plurality of input variables for the corresponding iteration.

6. The method of claim 5, wherein generating the overall variables contribution data comprises:

combining the plurality of variables contribution results for the plurality of iterations; and

determining a weighted average of the contribution for each of the plurality of input variables for the corresponding iteration.

7. The method of claim 1, wherein identifying the plurality of key input variables is based on a threshold, wherein the threshold is at least one of an area under curve statistic or a receiver operating characteristic curve statistic.

8. The method of claim 1, further comprising:

training a manufacturing-related prediction module using the plurality of key input variables.

9. A non-transitory computer readable storage medium including instructions that, when executed by a processing device, cause the processing device to perform a set of operations comprising:

executing a decision tree-based prediction model for a plurality of iterations using a plurality of input variables, wherein the plurality of iterations create corresponding rankings of the plurality of input variables;

generating overall variables contribution data using the rankings of the plurality of input variables; and

identifying, by the processing device, a plurality of key input variables based on the overall variables contribution data.

10. The non-transitory computer readable storage medium of claim 9, wherein generating the variables contribution data is based on at least one of a rules ensemble algorithm, a support vector machine algorithm, linear regression algorithm, or partial least squares algorithm.

11. The non-transitory computer readable storage medium of claim 9, wherein executing the decision tree-based prediction model comprises:

assigning a different weight to at least one of the plurality of input variables for the plurality of iterations, wherein the different weight is at least one of assigned using on at least one of an order of the plurality of input variable in a vector or assigned randomly.

12. The non-transitory computer readable storage medium of claim 9, wherein generating the overall variables contribution data comprises:

generating a plurality of variables contribution results corresponding to the plurality of iterations, the plurality of variables contribution results indicating a plurality of contributions corresponding to the plurality of input variables for the corresponding iteration;

combining the plurality of variables contribution results for the plurality of iterations; and

determining a weighted average of the contribution for each of the plurality of input variables for the corresponding iteration.

13. The non-transitory computer readable storage medium of claim 9, wherein identifying the plurality of key input variables is based on a threshold, wherein the threshold is at least one of an area under curve statistic or a receiver operating characteristic curve statistic.

14. The non-transitory computer readable storage medium of claim 9, the operations further comprising:

training a manufacturing-related prediction module using the plurality of key input variables.

15. A system comprising:

a memory to store a plurality of input variables; and

a processing device coupled to the memory to

execute a decision tree-based prediction model for a plurality of iterations using the plurality of input variables, wherein the plurality of iterations create corresponding rankings of the plurality of input variables;

generate overall variables contribution data using the rankings of the plurality of input variables; and

identify a plurality of key input variables based on the overall variables contribution data.

16. The system of claim 15, wherein generating the variables contribution data is based on at least one of a rules ensemble algorithm, a support vector machine algorithm, linear regression algorithm, or partial least squares algorithm.

17. The system of claim 15, wherein executing the decision tree-based prediction model comprises:

assigning a different weight to at least one of the plurality of input variables for the plurality of iterations, wherein the different weight is at least one of assigned using on at least one of an order of the plurality of input variable in a vector or assigned randomly.

18. The system of claim 15, wherein generating the overall variables contribution data comprises:

generating a plurality of variables contribution results corresponding to the plurality of iterations, the plurality of variables contribution results indicating a plurality of contributions corresponding to the plurality of input variables for the corresponding iteration;

combining the plurality of variables contribution results for the plurality of iterations; and

determining a weighted average of the contribution for each of the plurality of input variables for the corresponding iteration.

19. The system of claim 15, wherein identifying the plurality of key input variables is based on a threshold, wherein the threshold is at least one of an area under curve statistic or a receiver operating characteristic curve statistic.

20. The system of claim 15, wherein the processing device is further to:

train a manufacturing-related prediction module using the plurality of key input variables.