Back propagation neural network based big data analytics for a stock market challenge

Abstract

In this article we are presenting our methodology on solving a stock market challenge on predicting the intraday stock returns. We are presenting our complete approach on solving this challenge namely, the approaches to prepare the data from the unstructured data and the challenges on using back propagation neural network algorithm, namely the choice of activation function, learning rate and the number of neurons in the hidden layer. The validation of the approach is also presented demonstrating the effectiveness of back propagation neural network based model on predicting the stock returns. It was observed that the proposed algorithm was able to predict the stock returns with an maximum absolute error of and therefore the prediction is very close to the actual value.