Machine learning techniques for generating hybrid risk scores
Various embodiments of the present invention provide methods, apparatus, systems, computing devices, computing entities, and/or the like for performing risk score generation predictive data analysis. Certain embodiments of the present invention utilize systems, methods, and computer program products that risk score generation predictive data analysis by utilizing at least one of inferred hybrid risk score generation machine learning models and hybrid graph-based machine learning models.
1 . A computer-implemented method comprising:
receiving, by one or more processors, a hybrid risk score from an inferred hybrid risk score generation machine learning model based on a plurality of graph feature embedding data objects for a data object, wherein the inferred hybrid risk score generation machine learning model is trained by:
receiving, from a hybrid graph-based machine learning model, a set of inferred hybrid risk scores, wherein an inferred hybrid risk score of the set of inferred hybrid risk scores is generated based on a plurality of prior graph feature embedding data objects for a prior data object of a set of prior data objects,
determining a plurality of regressor variable values based on the plurality of prior graph feature embedding data objects for the prior data object,
training the inferred hybrid risk score generation machine learning model based on performing a set of genetic programming operations on the set of inferred hybrid risk scores for the set of prior data objects, wherein the plurality of regressor variable values and the set of inferred hybrid risk scores is provided as input to the set of genetic programming operations, and
storing the inferred hybrid risk score generation machine learning model in place of the hybrid graph-based machine learning model to reduce a runtime computation cost of the hybrid risk score;
determining, by the one or more processors, explanatory metadata based on the plurality of regressor variable values associated with the inferred hybrid risk score generation machine learning model in relation to the data object; and
providing, by the one or more processors, the hybrid risk score and the explanatory metadata to one or more client computing entities.
2 . The computer-implemented method of claim 1 , wherein the hybrid graph-based machine learning model comprises a plurality of graph-based machine learning models and an ensemble machine learning model.
3 . The computer-implemented method of claim 2 , wherein a graph-based machine learning model of the plurality of graph-based machine learning models is configured to process a prior graph feature embedding data object of the plurality of prior graph feature embedding data objects for the prior data object of the set of prior data objects to generate the inferred hybrid risk score of the set of inferred hybrid risk scores for the prior data object.
4 . The computer-implemented method of claim 2 , wherein the plurality of graph-based machine learning models comprises one or more graph convolutional neural network machine learning models.
5 . The computer-implemented method of claim 1 , wherein the plurality of graph feature embedding data objects comprises a genomic graph feature embedding data object.
6 . The computer-implemented method of claim 1 , wherein the plurality of graph feature embedding data objects comprises a behavioral graph feature embedding data object.
7 . The computer-implemented method of claim 1 , wherein the plurality of graph feature embedding data objects comprises a clinical graph feature embedding data object.
8 . The computer-implemented method of claim 1 , wherein the set of genetic programming operations comprises a set of symbolic regression operations that infer non-complex algebraic relationships between the plurality of prior graph feature embedding data objects.
9 . The computer-implemented method of claim 1 , wherein a prior graph feature embedding data object of the plurality of prior graph feature embedding data objects for a prior detect object is determined based on a risk tensor of a plurality of risk tensors for the prior data object.
10 . A system comprising:
one or more processors; and
at least one memory storing processor-executable instructions that, when executed by any one or more of the one or more processors, causes the one or more processors to perform operations comprising:
receive a hybrid risk score from an inferred hybrid risk score generation machine learning model based on a plurality of graph feature embedding data objects for a data object, wherein the inferred hybrid risk score generation machine learning model is trained by:
receiving, from a hybrid graph-based machine learning model, a set of inferred hybrid risk scores, wherein an inferred hybrid risk score of the set of inferred hybrid risk scores is generated based on a plurality of prior graph feature embedding data objects for a prior data object of a set of prior data objects,
determining a plurality of regressor variable values based on the plurality of prior graph feature embedding data objects for the prior data object,
training the inferred hybrid risk score generation machine learning model based on performing a set of genetic programming operations on the set of inferred hybrid risk scores for the set of prior data objects, wherein the plurality of regressor variable values and the set of inferred hybrid risk scores is provided as input to the set of genetic programming operations, and
storing the inferred hybrid risk score generation machine learning model in place of the hybrid graph-based machine learning model to reduce a runtime computation cost of the hybrid risk score;
determine explanatory metadata based on the plurality of regressor variable values associated with the inferred hybrid risk score generation machine learning model in relation to the data object; and
provide the hybrid risk score and the explanatory metadata to one or more client computing entities.
11 . The system of claim 10 , wherein the hybrid graph-based machine learning model comprises a plurality of graph-based machine learning models and an ensemble machine learning model.
12 . The system of claim 11 , wherein a graph-based machine learning model of the plurality of graph-based machine learning models is configured to process a prior graph feature embedding data object of the plurality of prior graph feature embedding data objects for the prior data object of the set of prior data objects to generate the inferred hybrid risk score of the set of inferred hybrid risk scores for the prior data object.
13 . The system of claim 11 , wherein the plurality of graph-based machine learning models comprises one or more graph convolutional neural network machine learning models.
14 . The system of claim 10 , wherein the plurality of graph feature embedding data objects comprises a genomic graph feature embedding data object.
15 . The system of claim 10 , wherein the plurality of graph feature embedding data objects comprises a behavioral graph feature embedding data object.
16 . The system of claim 10 , wherein the plurality of graph feature embedding data objects comprises a clinical graph feature embedding data object.
17 . The system of claim 10 , wherein the set of genetic programming operations comprises a set of symbolic regression operations that infer non-complex algebraic relationships between the plurality of prior graph feature embedding data objects.
18 . One or more non-transitory computer-readable storage media including instructions that, when executed by one or more processors, cause the one or more processors to:
receive a hybrid risk score from an inferred hybrid risk score generation machine learning model based on a plurality of graph feature embedding data objects for a data object, wherein the inferred hybrid risk score generation machine learning model is trained by:
receiving, from a hybrid graph-based machine learning model, a set of inferred hybrid risk scores, wherein an inferred hybrid risk score of the set of inferred hybrid risk scores is generated based on a plurality of prior graph feature embedding data objects for a prior data object of a set of prior data objects,
determining a plurality of regressor variable values based on the plurality of prior graph feature embedding data objects for the prior data object,
training the inferred hybrid risk score generation machine learning model based on performing a set of genetic programming operations on the set of inferred hybrid risk scores for the set of prior data objects, wherein the plurality of regressor variable values and the set of inferred hybrid risk scores is provided as input to the set of genetic programming operations, and
storing the inferred hybrid risk score generation machine learning model in place of the hybrid graph-based machine learning model to reduce a runtime computation cost of the hybrid risk score;
determine explanatory metadata based on the plurality of regressor variable values associated with the inferred hybrid risk score generation machine learning model in relation to the data object; and
provide the hybrid risk score and the explanatory metadata to one or more client computing entities.
19 . The one or more non-transitory computer-readable storage media of claim 18 , wherein the hybrid graph-based machine learning model comprises a plurality of graph-based machine learning models and an ensemble machine learning model.
20 . The one or more non-transitory computer-readable storage media of claim 19 , wherein a graph-based machine learning model of the plurality of graph-based machine learning models is configured to process a prior graph feature embedding data object of the plurality of prior graph feature embedding data objects for the prior data object of the set of prior data objects to generate the inferred hybrid risk score of the set of inferred hybrid risk scores for the prior data object.