MultiLayer ANNs: predicting the S&P 500 index

Abstract

Stock prediction with artificial neural network (ANN) models has been used extensively by researchers as it provides better results than other techniques. This paper presents an ANN approach to forecast the S&P 500 stock index price. First of all, an ANN-based variable selection model is presented. This model explores the relationship between some initial input variables and the closing price of the S&P 500 stock index. Furthermore, this research investigates how the training algorithm, as well as the number of neurons in the hidden layer and the distribution of the training data, affect the accuracy of the network.