Data model processing in machine learning using a reduced set of features

Patent metadata

Field	Value
Publication number	US-11429899-B2
Application number	US-202016863460-A
Country	US
Kind code	B2
Filing date	Apr 30, 2020
Priority date	Apr 30, 2020
Publication date	Aug 30, 2022
Grant date	Aug 30, 2022

Abstract

A computer system trains a predictive model. A plurality of subsets of features are selected from a dataset comprising a plurality of cases and controls and a plurality of features. Cases and controls are matched to select a plurality of case-control subsets for each subset of features, each case-control subset having similar values for the corresponding subset of features. For each case-control subset, a statistical significance of each feature of the plurality of features absent from the subset of features used to match the case-control subset is identified. A final subset of features is selected based on satisfying a statistical significance of each feature for the plurality of case-control subsets. A predictive model is trained using the final subset of features. Embodiments of the present invention further include a method and program product for training a predictive model in substantially the same manner described above.

First claim

The invention claimed is:

1. A computer-implemented method for training a predictive model comprising:
selecting, from a dataset comprising a plurality of cases and controls and a plurality of features, a plurality of subsets of features;
matching cases and controls to select a plurality of case-control subsets for each subset of features, each case-control subset having similar values for the corresponding subset of features;
identifying, for each case-control subset, a statistical significance of each feature of the plurality of features absent from the subset of features used to match the case-control subset;
selecting a final subset of features based on the statistical significance of each feature for the plurality of case-control subsets; and
training a predictive model using the final subset of features.

2. The computer-implemented method of claim 1 , further comprising:
applying the predictive model to predict outcomes.

3. The computer-implemented method of claim 1 , wherein selecting the final subset of features comprises:
determining a selection score for each feature of the plurality of features, wherein the selection score corresponds to a number of case-control subsets in which the statistical significance of the feature satisfies a significance threshold value; and
ranking the plurality of features by selection score to select the final subset of features having selection scores that satisfy a selection threshold value.

4. The computer-implemented method of claim 3 , wherein the significance threshold value comprises a probability score of the feature.

5. The computer-implemented method of claim 3 , wherein the selection threshold value comprises a percentage of case-control subsets in which the statistical significance of the feature satisfies the significance threshold value.

6. The computer-implemented method of claim 1 , further comprising:
evaluating the predictive model against a reference model to validate accuracy of the predictive model, wherein the reference model is trained using the dataset.

7. The computer-implemented method of claim 1 , wherein each case-control subset is matched according to propensity score matching with a caliper value and a case-control ratio value.

8. A computer system for training a predictive model, the computer system comprising:
one or more computer processors;
one or more computer readable storage media;
program instructions stored on the one or more computer readable storage media for execution by at least one of the one or more computer processors, the program instructions comprising instructions to:
select, from a dataset comprising a plurality of cases and controls and a plurality of features, a plurality of subsets of features;
match cases and controls to select a plurality of case-control subsets for each subset of features, each case-control subset having similar values for the corresponding subset of features;
identify, for each case-control subset, a statistical significance of each feature of the plurality of features absent from the subset of features used to match the case-control subset;
select a final subset of features based on the statistical significance of each feature for the plurality of case-control subsets; and
train a predictive model using the final subset of features.

9. The computer system of claim 8 , wherein the program instructions further comprise instructions to:
apply the predictive model to predict outcomes.

10. The computer system of claim 8 , wherein the program instructions to select the final subset of features comprise instructions to:
determine a selection score for each feature of the plurality of features, wherein the selection score corresponds to a number of case-control subsets in which the statistical significance of the feature satisfies a significance threshold value; and
rank the plurality of features by selection score to select the final subset of features having selection scores that satisfy a selection threshold value.

11. The computer system of claim 10 , wherein the significance threshold value comprises a probability score of the feature.

12. The computer system of claim 10 , wherein the selection threshold value comprises a percentage of case-control subsets in which the statistical significance of the feature satisfies the significance threshold value.

13. The computer system of claim 8 , wherein the program instructions further comprise instructions to:
evaluate the predictive model against a reference model to validate accuracy of the predictive model, wherein the reference model is trained using the dataset.

14. The computer system of claim 8 , wherein each case-control subset is matched according to propensity score matching with a caliper value and a case-control ratio value.

15. A computer program product for training a predictive model, the computer program product comprising one or more computer readable storage media collectively having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to:
select, from a dataset comprising a plurality of cases and controls and a plurality of features, a plurality of subsets of features;
match cases and controls to select a plurality of case-control subsets for each subset of features, each case-control subset having similar values for the corresponding subset of features;
identify, for each case-control subset, a statistical significance of each feature of the plurality of features absent from the subset of features used to match the case-control subset;
select a final subset of features based on the statistical significance of each feature for the plurality of case-control subsets; and
train a predictive model using the final subset of features.

16. The computer program product of claim 15 , wherein the program instructions further cause the computer to:
apply the predictive model to predict outcomes.

17. The computer program product of claim 15 , wherein the program instructions to select the final subset of features cause the computer to:
determine a selection score for each feature of the plurality of features, wherein the selection score corresponds to a number of case-control subsets in which the statistical significance of the feature satisfies a significance threshold value; and
rank the plurality of features by selection score to select the final subset of features having selection scores that satisfy a selection threshold value.

18. The computer program product of claim 17 , wherein the significance threshold value comprises a probability score of the feature.

19. The computer program product of claim 17 , wherein the selection threshold value comprises a percentage of case-control subsets in which the statistical significance of the feature satisfies the significance threshold value.

20. The computer program product of claim 15 , wherein the program instructions further cause the computer to:
evaluate the predictive model against a reference model to validate accuracy of the predictive model, wherein the reference model is trained using the dataset.

Assignees

• IBM
• Massachusetts Inst Technology

Inventors

• Kartoun Uri
• Severson Kristen
• Ng Kenney
• Myers Paul D
• Dai Wangzhi
• Stultz Collin M

Classifications

• G06N20/00 Primary

  Machine learning · CPC title

• G06N5/02

  Knowledge representation; Symbolic representation · CPC title