Forecasting Stock Returns using Evolutionary Artificial Neural Networks

Several models and techniques have been used to forecast stock returns. Some previous researches have applied Evolutionary Artificial Neural Networks to predict stocks prices. The most of previous researches have been concentrated on either predict stock indexes or trends rather than individual stock returns. In this paper, we use the adaptive EANNs to predict individual stock returns based on multivariate time series (AR with state variables) models. Comparing with the traditional Linear Regressions, the ANNs show promising results and our proposed EANNs can improve the performances of the ANNs as we expected even the improvements are slights