Interpretable feature discovery with grammar-based bayesian optimization
A method, a system, and a computer program product for generating an interpretable set of features. One or more search parameters and one or more constraints on one or more search parameters for searching data received from one or more data sources are defined. The data received from one or more data sources is searched using the defined search parameters and constraints. One or more first features are extracted from the searched data. The first features are associated with one or more predictive score values. The searching is repeated in response to receiving a feedback data responsive to the extracted first features. One or more second features resulting from the repeated searching are generated.
1. A computer implemented method, comprising:
defining, using at least one processor, one or more search parameters and one or more constraints on the one or more search parameters for searching data received from one or more data sources, and searching, using the defined one or more search parameters and one or more constraints, the data received from one or more data sources;
extracting, using the at least one processor, one or more first features from the searched data, the one or more first features being associated with one or more predictive score values;
repeating, using the at least one processor, the searching in response to receiving a feedback data responsive to the extracted one or more first features, and generating one or more second features resulting from the repeated searching, and
evaluating one or more of the first and second features using one or more objective functions, and performing the repeating based on the evaluating,
the one or more objective functions including at least one of the following: a function determining a stand-alone value of a binned first or second feature, a function determining an incremental value of a binned first or second feature, and any combination thereof.
2. The method according to claim 1 , wherein the searching is executed using a Bayesian search.
3. The method according to claim 1 , wherein the data includes at least one of the following: a transaction data, a time-series data, a tradeline data, a snapshot data, and any combination thereof.
4. The method according to claim 1 , wherein the one or more search parameters comprise at least one of the following: one or more data filter parameters, one or more data aggregation parameters, and any combination thereof.
5. The method according to claim 1 , wherein the one or more constraints include one or more constraining values associated with at least one respective search parameter in the one or more search parameters.
6. The method according to claim 1 , wherein least one of the first and second features are determined using one or more feature operators as a function of the one or more search parameters and the one or more constraints.
7. A system comprising:
at least one programmable processor; and
a non-transitory machine-readable medium storing instructions that, when executed by the at least one programmable processor, cause the at least one programmable processor to perform operations comprising:
defining, using at least one processor, one or more search parameters and one or more constraints on the one or more search parameters for searching data received from one or more data sources, and searching, using the defined one or more search parameters and one or more constraints, the data received from one or more data sources;
extracting, using the at least one processor, one or more first features from the searched data, the one or more first features being associated with one or more predictive score values;
repeating, using the at least one processor, the searching in response to receiving a feedback data responsive to the extracted one or more first features, and generating one or more second features resulting from the repeated searching, and
evaluating one or more of the first and second features using one or more objective functions, and performing the repeating based on the evaluating,
the one or more objective functions including at least one of the following: a function determining a stand-alone value of a binned first or second feature, a function determining an incremental value of a binned first or second feature, and any combination thereof.
8. The system according to claim 7 , wherein the searching is executed using a Bayesian search.
9. The system according to claim 7 , wherein the data includes at least one of the following: a transaction data, a time-series data, a tradeline data, a snapshot data, and any combination thereof.
10. The system according to claim 9 , wherein the one or more constraints include one or more constraining values associated with each at least one respective search parameter in the one or more search parameters.
11. The system according to claim 7 , wherein the one or more search parameters comprise at least one of the following: one or more data filter parameters, one or more data aggregation parameters, and any combination thereof.
12. The system according to claim 7 , wherein at least one of the first and second features are determined using one or more feature operators as a function of the one or more search parameters and the one or more constraints.
13. A computer program product comprising a non-transitory machine-readable medium storing instructions that, when executed by at least one programmable processor, cause the at least one programmable processor to perform operations comprising:
defining, using at least one processor, one or more search parameters and one or more constraints on the one or more search parameters for searching data received from one or more data sources, and searching, using the defined one or more search parameters and one or more constraints, the data received from one or more data sources;
extracting, using the at least one processor, one or more first features from the searched data, the one or more first features being associated with one or more predictive score values;
repeating, using the at least one processor, the searching in response to receiving a feedback data responsive to the extracted one or more first features, and generating one or more second features resulting from the repeated searching, and
evaluating one or more of the first and second features using one or more objective functions, and performing the repeating based on the evaluating,
the one or more objective functions including at least one of the following: a function determining a stand-alone value of a binned first or second feature, a function determining an incremental value of a binned first or second feature, and any combination thereof.
14. The computer program product according to claim 13 , wherein the searching is executed using a Bayesian search.
15. The computer program product according to claim 13 , wherein the data includes at least one of the following: a transaction data, a time-series data, a tradeline data, a snapshot data, and any combination thereof.
16. The computer program product according to claim 13 , wherein the one or more search parameters comprise at least one of the following: one or more data filter parameters, one or more data aggregation parameters, and any combination thereof.
17. The computer program product according to claim 13 , wherein the one or more constraints include one or more constraining values associated with at least one respective search parameter in the one or more search parameters.
18. The computer program product according to claim 13 , wherein least one of the first and second features are determined using one or more feature operators as a function of the one or more search parameters and the one or more constraints.