Reusable Knowledge from Symbolic Regression Classification

In this paper we generalize the well known regression method to fulfill supervised classification aiming to produce a learning model which best separates the class members of a labeled training set. The separation surface is represented by the level set of a model function and it is defined by the respective equation. The model is represented by mathematical formulas and composed of an optimum set of expressions of a given superset. We show that this property gives human experts additional insight in the application domain. Furthermore the representation in terms of mathematical formulas (e.g. the analytical model and its first and second derivative) adds additional value to the classifier and enables to answer questions which other classifier approaches cannot.