On the development of improved adaptive models for efficient prediction of stock indices using clonal-PSO (CPSO) and PSO techniques

The present paper introduces a new clonal particle swarm optimisation (CPSO) and PSO techniques to develop efficient adaptive forecasting models for short and long-term prediction of S&P 500 and DJIA stock indices. The basic structure of the models is an adaptive linear combiner whose weights are iteratively updated by PSO and CPSO-based learning rules. The technical indicators are computed from past stock indices and are used as input to the models. Using simulation study the prediction performances in terms of the convergence rate, the minimum mean square error (MSE), training time and the mean average percentage error (MAPE) of CPSO, PSO and GA-based models are obtained for all ranges of prediction. Comparison of these results demonstrates that the proposed CPSO and PSO-based models yield superior performance compared to the GA one. However the CPSO model provides the best performance compared to other two.