Ensemble with radial basis function neural networks for Casablanca stock market returns prediction

We present a radial basis function neural network (RBFNN) ensemble system (ES) to predict Casablanca Stock Exchange (CSE) returns based on its microstructure modeling. Its performance is compared to each RBFNN component and the conventional auto-regressive moving average (ARMA) process. Based on the mean of absolute errors (MAE) and mean of squared errors (MSE), the forecasting results showed that the RBFNNES outperformed each of its RBFNN components and also the traditional ARMA model. Our obtained results suggest that the proposed approach could be promising for CSE returns modeling and forecasting.