Decision-making system and method based on supervised learning

1. A computer-implemented method, comprising:

training a decision engine based on a labeled training dataset to generate a trained decision engine, the labeled training dataset containing N records corresponding to N entities respectively, each record comprising a value relating to a performance variable and a plurality of values corresponding to a set of attribute variables, wherein the trained decision engine comprises a Gradient Boosted Tree (GBT) ensemble of a plurality of regression trees and the training comprises:

constructing an initial regression tree based on minimization of a cost function associated with a root node corresponding to a select attribute variable and iteratively branching a training population into sub-populations corresponding to a set of leaf nodes based on remaining attribute variables;

determining a residual value and a corresponding loss function associated with the initial regression tree;

constructing a next regression tree based on the residual value and the loss function of the initial regression tree; and

iteratively generating subsequent regression trees based on a predecessor regression tree's residual value and corresponding loss function associated therewith, the iterative generation continuing until a predetermined number of regression trees are obtained as a fitted ensemble of the plurality of regression trees operable as the trained decision engine;

receiving, over a communications network, input values corresponding to a plurality of attributes associated with a subject entity, the input values acquired from a user equipment (UE) device operated by the subject entity;

determining, by the trained decision engine, a decision score for the subject entity responsive to the input values;

decomposing the decision score into a set of linearly decomposed constituent components corresponding to the attributes of the subject entity;

uniquely identifying respective portions of the decision score contributed by the corresponding attributes of the subject entity; and

generating an action report with respect to the decision score using only a selected subset of the linearly decomposed constituent components, the selected subset of the linearly decomposed constituent components objectively determined based on respective contributions to the decision score.

2. The computer-implemented method as recited in claim 1 , further comprising generating the action report regardless of an adverse decision based on the decision score.

3. The computer-implemented method as recited in claim 1 , wherein the selected subset of the linearly decomposed constituent components are determined based on the attributes that contribute most to the decision score.

4. The computer-implemented method as recited in claim 1 , further comprising determining that the attributes of the subject entity corresponding to the selected subset of the linearly decomposed constituent components are compliant with respect to a set of regulatory compliance rules.

5. The computer-implemented method as recited in claim 1 , further comprising transmitting the action report to at least one of the subject entity, a governmental agency, a financial institution, and a third-party entity.

6. The computer-implemented method as recited in claim 1 , wherein the next regression tree and the subsequent regression trees are generated using a gradient descent process.

7. The computer-implemented method as recited in claim 1 , further comprising:

selecting a subset of attribute variables; and

constructing the next regression tree and the subsequent regression trees using only the selected subset of attribute variables.

8. The computer-implemented method as recited in claim 7 , wherein the subset of attribute variables are randomly selected from the set of attribute variables.

9. The computer-implemented method as recited in claim 1 , wherein the set of attribute variables comprise one or more socio-economic variables, demographic variables, medical history variables, financial history variables, and variables based on social media network profiles for the N entities.

10. The computer-implemented method as recited in claim 1 , further comprising storing only the selected subset of the linearly decomposed constituent components.

11. An apparatus, comprising:

one or more processors;

one or more persistent memory modules coupled to the one or more processors, the one or more persistent memory modules having program instructions stored thereon which, when executed by the one or more processors, are configured to perform following acts:

training a decision engine based on a labeled training dataset to generate a trained decision engine, the labeled training dataset containing N records corresponding to N entities respectively, each record comprising a value relating to a performance variable and a plurality of values corresponding to a set of attribute variables, wherein the trained decision engine comprises a Gradient Boosted Tree (GBT) ensemble of a plurality of regression trees and the training comprises:

constructing an initial regression tree based on minimization of a cost function associated with a root node corresponding to a select attribute variable and iteratively branching a training population into sub-populations corresponding to a set of leaf nodes based on remaining attribute variables;

determining a residual value and a corresponding loss function associated with the initial regression tree;

constructing a next regression tree based on the residual value and the loss function of the initial regression tree; and

iteratively generating subsequent regression trees based on a predecessor regression tree's residual value and corresponding loss function associated therewith, the iterative generation continuing until a predetermined number of regression trees are obtained as a fitted ensemble of the plurality of regression trees operable as the trained decision engine;

receiving, over a communications network, input values corresponding to a plurality of attributes associated with a subject entity, the input values acquired from a user equipment (UE) device operated by the subject entity;

determining, by the trained decision engine, a decision score for the subject entity responsive to the input values;

decomposing the decision score into a set of linearly decomposed constituent components corresponding to the attributes of the subject entity; and

uniquely identifying respective portions of the decision score contributed by the corresponding attributes of the subject entity; and

a report generator configured to generate an action report with respect to the decision score using only a selected subset of the linearly decomposed constituent components, the selected subset of the linearly decomposed constituent components objectively determined based on respective contributions to the decision score.

12. The apparatus as recited in claim 11 , wherein the report generator is configured to generate the action report regardless of an adverse decision based on the decision score.

13. The apparatus as recited in claim 11 , wherein the selected subset of the linearly decomposed constituent components are determined based on the attributes that contribute most to the decision score.

14. The apparatus as recited in claim 11 , wherein the program instructions include instructions for determining that the attributes of the subject entity corresponding to the selected subset of the linearly decomposed constituent components are compliant with respect to a set of regulatory compliance rules.

15. The apparatus as recited in claim 11 , wherein the program instructions include instructions for transmitting the action report to at least one of the subject entity, a governmental agency, a financial institution, and a third-party entity.

16. The apparatus as recited in claim 11 , wherein the program instructions further comprise instructions for generating the next regression tree and the subsequent regression trees using a gradient descent process.

17. The apparatus as recited in claim 11 , wherein the program instructions further comprise instructions for:

selecting a subset of attribute variables; and

constructing the next regression tree and the subsequent regression trees using only the selected subset of attribute variables.

18. The apparatus as recited in claim 17 , wherein the subset of attribute variables are randomly selected from the set of attribute variables.

19. The apparatus as recited in claim 11 , wherein the set of attribute variables comprise one or more socio-economic variables, demographic variables, medical history variables, financial history variables, and variables based on social media network profiles for the N entities.

20. The apparatus as recited in claim 11 , further comprising a storage unit configured to store only the selected subset of the linearly decomposed constituent components.