Input feature significance identification based on batches of prediction

A computer system for identifying input feature significance for a machine learning model, the computer system comprising:

• a bus system;
• a storage device connected to the bus system, wherein the storage device stores program instructions;
• and
• a processor connected to the bus system, wherein the processor executes the program instructions to:
• slice a set of scoring payload data corresponding to a set of input features of the machine learning model into a predefined number of batches to form a sliced set of scoring payload data;
• generate, using the sliced set of scoring payload data, a correlation coefficient matrix for each respective input feature within the set of input features of the machine learning model for each batch in the predefined number of batches based on input of each particular input feature into the machine learning model and a corresponding output from the machine learning model for each particular input feature indicating all combinations of input features and corresponding machine learning model outputs per batch;
• extract a correlation coefficient value from the correlation coefficient matrix for each particular input feature and the corresponding output from the machine learning model per batch;
• identify a set of significant input features having a corresponding correlation coefficient value greater than or equal to a predefined correlation coefficient threshold level per batch;
• and
• perform a set of action steps automatically regarding the set of significant input features per batch.

A computer program product for identifying input feature significance for a machine learning model, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform a method of:

• slicing a set of scoring payload data corresponding to a set of input features of the machine learning model into a predefined number of batches to form a sliced set of scoring payload data;
• generating, using the sliced set of scoring payload data, a correlation coefficient matrix for each respective input feature within the set of input features of the machine learning model for each batch in the predefined number of batches based on input of each particular input feature into the machine learning model and a corresponding output from the machine learning model for each particular input feature indicating all combinations of input features and corresponding machine learning model outputs per batch;
• extracting a correlation coefficient value from the correlation coefficient matrix for each particular input feature and the corresponding output from the machine learning model per batch;
• identifying a set of significant input features having a corresponding correlation coefficient value greater than or equal to a predefined correlation coefficient threshold level per batch;
• and
• performing a set of action steps automatically regarding the set of significant input features per batch.