Log in

Genetic Programming of Fuzzy Logic Production Rules with Application to Financial Trading.

John Koza has demonstrated that a form of machine learning can be constructed  by using the techniques of Genetic Programming using LISP statements. We  describe here an extension to this principle using Fuzzy Logic sets and  operations instead of LISP expressions. We show that Genetic programming can  be used to generate trees of fuzzy logic statements, the evaluation of which 
optimise some external process, in our example financial trading. We also show that these trees can be simply converted to natural language rules, and that these rules are easily comprehended by a lay audience. This clarity of internal function can be compared to “Black Box” non-parametric modelling techniques such as Neural Networks. We then show that even with minimal data preparation the technique produces rules with good out of sample performance on a range of different financial instruments.