Optimal Combination of Trading Rules Using Neural Networks


  •  Subrata Mitra    

Abstract

A large number of trading rules based on technical analysis of prices are being used by investing community for generating trading signals for short term investments. As profitability of these trading rules vary, it is not easy to judge which particular rule really ‘works’. Instead of a single trading rule, combination of rules are likely to offer the portfolio benefits of better risk adjusted return and hence, an experiment is carried out to combine signals generated from of moving averages of different window size using an artificial neural network.  It is observed that the risk adjusted performance measure of the artificial neural network based trading model is better than that of simple ‘Buy and Hold’ strategy.


This work is licensed under a Creative Commons Attribution 4.0 License.