Neuro-Fuzzy and Particle Swarm optimization based Hybrid Approach for Stock Price Forecasting

This study, investigates the applicability of adaptive network based fuzzy inference system (ANFIS) and discrete particle swarm optimization (DPSO) based hybrid approach for forecasting the close price of two major stock indices of India viz. Nifty 50 and Sensex using optimum number of technical indicators. The primary focus of this work is to create a systematic approach based on swarm intelligence based neuro-fuzzy model where DPSO attempts to search automatically the optimum number of features to overcome overfitting problem and subtractive clustering determine the compact rule-base of neuro-fuzzy system. Experimental results reveal that the proposed hybrid model shows the superior forecasting performance in comparison with the model uses complete set of features.