Enhancing silent features with adversarial networks for improved model versions

1 . A computer-implemented method, comprising operations for:

obtaining input features of a first number of features;

identifying hidden features of a second number of new features;

performing quantum feature importance scoring to assign an importance score to each of the input features and to each of the hidden features, wherein the quantum feature importance scoring uses a permutation importance technique;

identifying silent features as the hidden features with the importance score below a first threshold;

identifying important features as the input features with the importance score above a second threshold and the hidden features with the importance score above the second threshold;

building a silent feature model using the silent features;

building an important feature model using the important features;

integrating the silent features and the important features into a single model by building an ensemble model with the silent feature model and the important feature model; and

using the ensemble model to generate one or more predictions and one or more prescriptions comprising one or more recommendations.

2 . The computer-implemented method of claim 1 , wherein the input features include one or more stored features and one or more features identified by web scraping.

3 . The computer-implemented method of claim 1 , wherein the hidden features are identified using a conditional Generative Adversarial Network (cGAN) to generate the second number of the new features.

4 . The computer-implemented method of claim 1 , wherein operations within a Quantum Model Store build the silent feature model and the important feature model.

5 . The computer-implemented method of claim 1 , further comprising operations for:

iteratively creating versions of the silent feature model until a version of the silent feature model is selected for storage based on statistics generated for the silent feature model.

6 . The computer-implemented method of claim 1 , further comprising operations for:

storing one or more versions of the silent feature model and one or more versions of the important feature model in a Quantum Model Store in a computer located in a cloud.

7 . The computer-implemented method of claim 1 , further comprising operations for:

normalizing the silent features; and

scaling the silent features.

8 . A computer program product, the computer program product comprising a computer readable storage medium having program code embodied therewith, the program code executable by at least one processor to perform operations for:

obtaining input features of a first number of features;

identifying hidden features of a second number of new features;

performing quantum feature importance scoring to assign an importance score to each of the input features and to each of the hidden features, wherein the quantum feature importance scoring uses a permutation importance technique;

identifying silent features as the hidden features with the importance score below a first threshold;

identifying important features as the input features with the importance score above a second threshold and the hidden features with the importance score above second threshold;

building a silent feature model using the silent features;

building an important feature model using the important features;

integrating the silent features and the important features into a single model by building an ensemble model with the silent feature model and the important feature model; and

using the ensemble model to generate one or more predictions and one or more prescriptions comprising one or more recommendations.

9 . The computer program product of claim 8 , wherein the input features include one or more stored features and one or more features identified by web scraping.

10 . The computer program product of claim 8 , wherein the hidden features are identified using a conditional Generative Adversarial Network (cGAN) to generate the second number of the new features.

11 . The computer program product of claim 8 , wherein operations within a Quantum Model Store build the silent feature model and the important feature model.

12 . The computer program product of claim 8 , wherein the program code is executable by the at least one processor to perform operations for:

iteratively creating versions of the silent feature model until a version of the silent feature model is selected for storage based on statistics generated for the silent feature model.

13 . The computer program product of claim 8 , wherein the program code is executable by the at least one processor to perform operations for:

storing one or more versions of the silent feature model and one or more versions of the important feature model in a Quantum Model Store in a computer located in a cloud.

14 . The computer program product of claim 8 , wherein the program code is executable by the at least one processor to perform operations for:

normalizing the silent features; and

scaling the silent features.

15 . A computer system, comprising:

one or more processors, one or more computer-readable memories and one or more computer-readable, tangible storage devices; and

program instructions, stored on at least one of the one or more computer-readable, tangible storage devices for execution by at least one of the one or more processors via at least one of the one or more computer-readable memories, to perform operations comprising:

obtaining input features of a first number of features;

identifying hidden features of a second number of new features;

performing quantum feature importance scoring to assign an importance score to each of the input features and to each of the hidden features, wherein the quantum feature importance scoring uses a permutation importance technique;

identifying silent features as the hidden features with the importance score below a first threshold;

identifying important features as the input features with the importance score above a second threshold and the hidden features with the importance score above the second threshold;

building a silent feature model using the silent features;

building an important feature model using the important features;

integrating the silent features and the important features into a single model by building an ensemble model with the silent feature model and the important feature model; and

using the ensemble model to generate one or more predictions and one or more prescriptions comprising one or more recommendations.

16 . The computer system of claim 15 , wherein the input features include one or more stored features and one or more features identified by web scraping.

17 . The computer system of claim 15 , wherein the hidden features are identified using a conditional Generative Adversarial Network (cGAN) to generate the second number of the new features.

18 . The computer system of claim 15 , wherein operations within a Quantum Model Store build the silent feature model and the important feature model.

19 . The computer system of claim 15 , wherein the operations further comprise:

iteratively creating versions of the silent feature model until a version of the silent feature model is selected for storage based on statistics generated for the silent feature model.

20 . The computer system of claim 15 , wherein the operations further comprise:

storing one or more versions of the silent feature model and one or more versions of the important feature model in a Quantum Model Store in a computer located in a cloud.