Using hyperheuristics under a GP framework for financial forecasting

  • Authors:

  • Michael Kampouridis;Edward Tsang

  • Affiliations:

  • School of Computer Science and Electronic Engineering, University of Essex, UK;Centre for Computational Finance and Economic Agents, University of Essex, UK

  • Venue:
  • LION'05 Proceedings of the 5th international conference on Learning and Intelligent Optimization
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Hyperheuristics have successfully been used in the past for a number of search and optimization problems. To the best of our knowledge, they have not been used for financial forecasting. In this paper we use a simple hyperheuristics framework to investigate whether we can improve the performance of a financial forecasting tool called EDDIE 8. EDDIE 8 allows the GP (Genetic Programming) to search in the search space of indicators for solutions, instead of using pre-specified ones; as a result, its search area is quite big and sometimes solutions can be missed due to ineffective search. We thus use two different heuristics and two different mutators combined under a simple hyperheuristics framework. We run experiments under five datasets from FTSE 100 and discover that on average, the new version can return improved solutions. In addition, the rate of missing opportunities reaches it's minimum value, under all datasets tested in this paper. This is a very important finding, because it indicates that thanks to the hyperheuristics EDDIE 8 has the potential of missing less forecasting opportunities. Finally, results suggest that thanks to the introduction of hyperheuristics, the search has become more effective and more areas of the space have been explored.