Robust technical trading strategies using GP for algorithmic portfolio selection

Abstract

This paper presents a Robust Genetic Programming approach for discovering profitable trading rules which are used to manage a portfolio of stocks from the Spanish market. The investigated method is used to determine potential buy, sell conditions for stocks, aiming to yield robust solutions able to withstand extreme market conditions, while producing high returns at a minimal risk. One of the biggest challenges GP evolved solutions face is over-fitting. GP trading rules need to have similar performance when tested with new data in order to be deployed in a real situation. We explore a random sampling method (RSFGP) which instead of calculating the fitness over the whole dataset, calculates it on randomly selected segments of it. This method shows improved robustness and out-of-sample results compared to standard genetic programming and a volatility adjusted fitness. Trading strategies (TS) are evolved using financial metrics like the volatility, CAPM alpha and beta, and the Sharpe ratio alongside other Technical Indicators (TI) to find the best investment strategy. These strategies are evaluated in using 21 of the most liquid stocks of the Spanish market. The achieved results clearly outperform both the Buy&Hold. Additionally, the solutions obtained with the training data during the experiments clearly show during testing robustness to step market declines as seen in the European sovereign debt experienced recently in Spain. In this paper the solutions learned where able to operate for prolonged periods, which demonstrated the validity and robustness of the rules learned, which are able to operate continuously and with minimal human intervention. To sum up, the developed method is able to evolve TSs suitable for all market conditions with promising results, which suggests great potential in the method generalization capabilities. The use of the financial metrics alongside popular TI enables the system to increase the stock return while proving resilient through time. The GP system is able to cope with different types of markets achieving a portfolio return of slightly higher than 30% for the period of 20092013 in the Spanish market, in a period that includes the sovereign debt crisis.

Dynamic portfolio insurance strategy: a robust machine learning approach

“…Note that since the risk variables have different ranges of values, they are normalized to the same range before being used in RGP. For the function set in this paper, we use the four commonly used arithmetic operators as follows (Berutich, López, Luna, & Quintana, 2016):…”

Index tracking optimization with cardinality constraint: a performance comparison of genetic algorithms and tabu search heuristics

“…Two other papers have recently been published by Aguilar-Rivera et al [1] and Berutich et al [8] dealing with evolutionary algorithms, stock markets and portfolio selection.…”

Smoothed Heikin-Ashi Algorithms Optimized for Automated Trading Systems

“…Different analytical 199 Economic Informatics Doctoral School, Bucharest Academy of Economic Studies, 11th Tache Ionesc Str., Bucharest, Romania This paper was co-financed by the Bucharest Academy of Economic Studies during the PhD program I trading models that can be automated for the financial markets are presented by Larry Connors in [6] and [7] with proved reliable results. Using the genetic based algorithms, interesting researches are presented by Yong Hu [8] and José Manuel Berutich [9]. Hybrid trading strategies are also developed in the last time; a hybrid system is presented by Youngmin Kim [10].…”