Improving return using risk-return adjustment and incremental training in technical trading rules with GAPs

Abstract

The principle objective of this paper is to obtain trading rules with a low risk level which are also capable of obtaining high returns. To that purpose a methodology has been defined, based on the design of a genetic algorithm GAP and an incremental training technique adapted to the learning of series of stock market values. The GAP technique consists in a fusion of GP and GA. In GAP a chromosome is composed of a tree with language operators and a vector with numeric values. The GAP algorithm implements the automatic search for trading rules taking as objectives of the training both the optimization of the return obtained and the minimization of the assumed risk. In order to diminish high over-fitting, a technique of incremental training has been used. Applying the proposed methodology, rules have been obtained for a period of eight years of the S&P500 index. The achieved adjustment of the relation return-risk has generated rules with returns very superior in the testing period to those obtained applying habitual methodologies and even clearly superior to Buy&Hold. Insert your abstract here. Include keywords, PACS and mathematical subject classification numbers as needed.