GPTIPS: An Open Source Genetic Programming Toolbox For Multigene Symbolic Regression 论文

摘要

Abstract — In this contribution GPTIPS, a free, open source MATLAB toolbox for performing symbolic regression by genetic programming (GP) is introduced. GPTIPS is specifically designed to evolve mathematical models of predictor response data that are “multigene ” in nature, i.e. linear combinations of low order non-linear transformations of the input variables. The functionality of GPTIPS is demonstrated by using it to generate an accurate, compact QSAR (quantitative structure activity relationship) model of existing toxicity data in order to predict the toxicity of chemical compounds. It is shown that the low-order “multigene ” GP methods implemented by GPTIPS can provide a useful alternative, as well as a complementary approach, to currently accepted empirical modelling and data analysis techniques. GPTIPS and documentation is available for download at