Title: A hybrid CRO-based FLANN for financial stock market forecasting

Authors: Soumya Ranjan Sahu; Himansu Sekhar Behera

Addresses: Department of Computer Science and Engineering and Information Technology, Veer Surendra Sai University of Technology (VSSUT), Burla, Sambalpur-768018, Odisha, India ' Department of Computer Science and Engineering and Information Technology, Veer Surendra Sai University of Technology (VSSUT), Burla, Sambalpur-768018, Odisha, India

Abstract: Stock market prediction has been a challenging area of research due to the dynamically changing stock market values. Many research works have been carried out to predict the stock market using different kinds of neural network model and different learning and optimisation technique. In this work a trigonometric FLANN model has been proposed for the forecasting of four different stock indices; BSE-S&P100, BSE-S&P500, NASDAQ and S&P500. A hybridised learning algorithm of least mean square (LMS) and chemical reaction optimisation (CRO) is used to train the model. Root mean square error (RMSE) is considered as the cost function for training phase and during the testing phase means average percentage error (MAPE) is used to evaluate the models. The results obtained are compared with that of the FLANN-LMS model. The experimental result shows that FLANN-CRO-LMS model performs significantly better than FLANN-LMS model.

Keywords: financial forecasting; stock markets; functional link ANNs; artificial neural networks; FLANN; chemical reaction optimisation; CRO; least mean squares; LMS; stock market prediction.

DOI: 10.1504/IJDATS.2016.079059

International Journal of Data Analysis Techniques and Strategies, 2016 Vol.8 No.3, pp.261 - 279

Received: 05 Mar 2014
Accepted: 29 Sep 2014
Published online: 12 Sep 2016 *

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