1. Introduction

For the Borrower Screening product, the variables are derived from our proprietary raw data through internal feature engineering and calculations, so we are unable to disclose the detailed definitions of individual variables externally.

Recommendations

Typically, we recommend that clients first obtain a sample of the test variables and evaluate them against their own labels by calculating metrics such as KS (Kolmogorov–Smirnov) values.

Features that demonstrate strong discriminatory power can be incorporated directly into business rules or strategies, while those with lower discriminatory power can be further evaluated as part of a predictive modeling process.