Stock trading with random forests, trend detection tests and force index volume indicators

Abstract

The goal of this paper is to investigate if the strong machine learning technique is able to retrieve information from past prices and predict price movements and future trends. The architecture of the system with the on-line adaptation ability to non-stationary two dimensional mixed Black-Scholes Markov time series model is presented. The methodology of investment strategies performance verification is also proposed. © 2013 Springer-Verlag.