Metaheuristics application on a financial forecasting problem

摘要: EDDIE is a Genetic Programming (GP) tool, which used to tackle problems in the field of financial forecasting. The novelty its grammar, allows GP look space technical analysis indicators, instead using prespecified ones, as it normally happens literature. advantage this that not constrained use indicators; instead, thanks can choose any indicators within pre-defined range, leading new solutions might have never been discovered before. However, disadvantage above approach algorithm's search dramatically larger, and result good sometimes be missed due ineffective search. This paper presents an attempt deal with issue by applying three different meta-heuristics, namely Simulated Annealing, Tabu Search, Guided Local Search. Results show performance significantly improves, thus making combination meta-heuristics effective forecasting approach.