Stock Prediction Research Based On MEA-BP Model

Stock market prediction is one of the hot topics in Finance Research. Due to many influenced factors, the stock market shows complex nonlinearity, which means that the traditional methods based on linear analysis are inadequate. The neural network can achieve the nonlinear mapping between arbitrary spaces, thus are favored by many scholar. Although there are many stock prediction researches based on BP or GA-BP neural network, few researches employ MEA-BP neural network. In this paper, we first selects 8 basic indexes of stock index to make the case study by the BP neural network with the historical data of the Shanghai Composite Index. The results show that on the one hand, although the BP neural network can predict the change of stock market trend, the deviation from the true value is also large; on the other hand, the GA-BP neural network has high precision, but its convergence speed is slow. In view of this point, we further employ Mind Evolutionary Algorithm (MEA) to optimize the connection weights and thresholds of BP neural network, and compare with GA-BP neural network. The results show that not only is the prediction accuracy of MEA-BP neural network higher, but also the convergence speed is faster.