Stock Price Prediction Using Genetic Neuro-Fuzzy Model

The need to accurately predict the price of stock before trading is important as it will minimize loss and maximize profit. A lot of approaches have been used to do this but the results obtained have not been satisfactory. This paper therefore presents a hybrid of genetic algorithm and the adaptive neuro-fuzzy systems for stock price prediction. A neuro-fuzzy model was optimized using Genetic Algorithm which is an evolutionary approach. The model was tested using stock dataset of First bank Nigeria PLC. The model was trained using 2001 data items consisting of 3 attributes obtained during feature selection. The model?s parameter obtained from the training are then saved. For testing, 228 data items are used to test the model. Out of this, 172 are classified correctly while 56 were misclassified. The model was compared with a neural networks model, a decision tree model and a neuro-fuzzy model. The model outperforms these models by having the lowest mean square error.