Portfolio allocation using XCS experts in technical analysis, market conditions and options market

Abstract

Schulenburg [15] first proposed the idea to model different trader types by supplying different input information sets to a group of homogenous LCS agent. Gershoff [12] investigated this idea further with XCS agent. This paper takes an extra step to build a trading system that not only adopts the multi-XCS agent idea, but also utilizes knowledge from discretization theory, modern portfolio theory, options theory and methods of combining multiple models. In comparison to previous work, a wider range of input data were used including technical analysis, general market conditions and options market conditions. Secondly, quantization of continuous financial series was achieved using entropy-based discretization and histogram equalization. Thirdly, subtle investment strategies can now be generated as a result of taking stock price magnitude into account. Finally, multiple agents' predictions were combined using a variant of stacking. Empirical results show the best-performing XCS agents always outclass benchmark agents in every stock examined. Variance is reduced after combining predictions from multiple models. The technical analysis XCS agent was able to replicate a well known technical trading rule widely used in the 60s.

Knowledge Representation in Learning Classifier Systems: A Review

“…The year 1995 is marked a milestone in LCS researches due to the appearance of the most successful and popular Michigan style LCS named XCS. Recent analysis and researches (Wong and Schulenburg, 2007;Studley and Bull, 2005;Vargas et al, 2004) have shown that XCS has an especial architecture that makes it a competitive learning system in solving complex problems. Its architecture ingredients provide some promising properties such as its online learning capability, its noise robustness, the generality in the learning mechanism, and its continuous adaptation.…”

“…From 1995 to now, the applicability of XCS has been extended to a wide range of applications including computational economics (Schulenburg and Ross, 1999;Wong and Schulenburg, 2007), classification and data mining (Bull et al, 2008), autonomous robotics (Patel and Dorigo, 1994; Studley, 2005;Studley and Bull, 2005), power distribution network (Vargas et al, 2004), traffic light control (Bull et al, 2004), function approximation tasks (Wilson, 2002) and many more (Bull, 2004;Lanzi et al, 2000;Bull et al, 2008). So, the current LCS researches are very diverse and pervasive and done better than ever.…”

Generation of rule-based adaptive strategies for a collaborative virtual simulation environment

“…They have been studied in the domain of Data Mining [32], economic simulations [33], medical data analysis [34] and personal software agents [35]. In video games domain, [36] applied FXCS in a cooperative game environment, and [37] developed an action selector for a MMORPG.…”