Binning for Nonlinear Modeling

A method, comprising:

• binning variables in the training data set into bins;
• binning variables in a validation data set into the bins;
• generating a decision tree based on the bins, the binned variables from the training data set, and the binned variables from the validation data set;
• sorting the leaf nodes in the decision tree;
• merging leaf nodes of the decision tree until a weight of evidence associated with the training data set and a weight of evidence associated with the validation data set are both monotonic in the sorted leaf nodes;
• and
• determining bins for a data model from the merged leaf nodes.

A non-transitory machine-readable medium having stored thereon machine-readable instructions executable to cause a machine to perform operations comprising:

• determining bins for a training data set;
• binning variables in the training data set into the bins;
• binning variables in a validation data set into the bins;
• generating a first decision tree using the bins and the binned variables from the training data set;
• generating a second decision tree using the bins and the binned variables from the variable data set;
• merging a first decision tree and a second decision tree into a third decision tree;
• merging leaf nodes of the third decision tree until a weight of evidence associated with the training data set and a weight of evidence associated with the validation data set are both monotonic;
• and
• determining bins for a data model from the merged leaf nodes.