Stock Price Prediction Based on ARIMA-RNN Combined Model

Abstract

This study proposes a multivariate Hamacher-adaptive neuro-fuzzy inference system (ANFIS) method for price modeling and forecasting in the stock market in China using subtraction-fuzzy clustering algorithm, multivariate Hamacher operator, and ANFIS. First, the multivariate Hamacher operator is combined with ANFIS to modify the membership measure mechanism and the rule parameter-updating mechanism of ANFIS rules. A multivariate Hamacher-ANFIS model based on subtraction-fuzzy clustering is then established. This study chooses five stocks with the largest total market capitalization from the Shanghai and Shenzhen stock markets to calculate their historical volatility in the same period and obtain the weight of model forecasting performance. Finally, model parameters are initialized using the subtractive-fuzzy clustering algorithm. Fivefold cross-checking is performed for each data set, and the integrated R2 value of the model with respect to the test set is calculated from the previously obtained weights t. Experimental results show that the model improves the learning capability of the complex objective function and the prediction accuracy of the stock price compared with existing methods.