Extracting Nonlinear Dependencies: An Easy, Automatic Method

Extracting nonlinear dependencies from the everyclient’s database is the key to successful data mining, ergo, successful predictive modeling. A nonlinear dependency is not the everyday linear combination of important predictor variables defining in part the standard statistical regression model. A nonlinear dependency maybe, say, Sine [{(X2 * (X1 – X2)} * (X1 – X2)]. But, how does one find such a nonlinear dependency? There are many articles on the topic, but no practical solution exists. More importantly, if any solution does exist, no commercial software appears to be readily available. Wrong! The purpose of this article is to present the GenIQ Model (software) as an easy, automatic method for extracting nonlinear dependencies, which when used as input into any predictive modeling approach will outdo any statistical linear method.