Terminology
- ATS (Adaptive Transformation Score) is a transformation based on a flexible, non-parametric regression and provides an easy-to-view visualization of the relationship between target variables and the predictors (explanatory variables). The ATS function has better predictive power than traditional linear or non-linear parametric transformations. It utilizes data better than optimal binning and other methods that group continuous variables in mutually exclusive buckets. Due to its flexibility of the transformation, it can generates smooth functions, hence eliminating time consuming manual smoothing techniques.