Title: An interactive tool for the stock market research using recursive neural networks

Authors: Michal Trna; Víctor Giménez-Martínez

Addresses: Departamento de Matemática Aplicada, Facultad de Informática, Universidad Politécnica de Madrid, 28660, Boadilla del Monte, Madrid, Spain. ' Departamento de Matemática Aplicada, Facultad de Informática, Universidad Politécnica de Madrid, 28660, Boadilla del Monte, Madrid, Spain

Abstract: The focus of this paper is to describe an application of a specific type of recurrent neural networks (RNN) in the domain of the technical analysis of the stock market. Said RNN is used as a tool that provides statistical information based on the historical time series processed in the training stage. In order to address certain questions of interest from the field of technical analysis of the financial markets, we introduce a set of constraints on the convergence of the state vector during the process of retrieval. Such approach retains its high computational efficiency notable in the unmodified version of the network, while it also provides an intuitive visualisation of the process of training and retrieval and thus facilitates an in-depth analysis of the system. We also present the pilot applications of exploiting information extracted from given time series to support an informed decision in an environment of real-time application.

Keywords: graph theory; recursive neural networks; RNN; market analysis; portfolio analysis; intelligence paradigm; interactive tool; stock markets.

DOI: 10.1504/IJAIP.2012.048140

International Journal of Advanced Intelligence Paradigms, 2012 Vol.4 No.2, pp.103 - 119

Published online: 23 Aug 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article