HEterogenous RAnking CummuLativE System - HERACLES

Optimal portfolio selection can be supported by the new trends in Artificial Intelligence, Deep Learning, and Portfolio Theory, examined in 4 neural system groups. The novel HEterogenous RAnking CumuLativE System – HERACLES, implements the optimal classifier that is discovered on the portfolio selection, to address key aspects of literature and the markets such as, the: i) Fractal Market Hypothesis (FMH) patterns in the Entropy of returns affected by the Chaotic Dynamics and following Kappa Distributions, ii) behavioral performance of investors, iii) detection of an optimal classifier on 1 Voted Perceptron (VP), 46 Multilayer Perceptrons (MLPs), 66 Modulars (MDNs), 60 Self-Organized Feature Maps models (SOFMs), in Neuro or/and Hybrid Neural-Genetic systems to optimise investments. HERACLES acts as a dragon-slayer to prevent bubbles in the markets.