Research on Economic and Financial Data Analysis Method Based on Genetic Algorithm

Abstract

The stock market study is the key point of the current economic exploration. As a relatively complex system in the development of financial economy in our country, green bamboo is not a small blood. Neural network is famous for its strong fitting ability of nonlinear system. Using neural network model to study stock prediction and putting forward genetic algorithm can effectively solve the problem of putting forward appropriate variable selection criteria which can not be solved by single neural network. The selection of variables based on genetic algorithm can optimize all the variables that affect the stock price globally, so as to deal with the selection of input variables of neural network. Based on the analysis of existing research literature, this paper proposes a new applicability function, which not only analyzes the prediction error, but also studies the number of variables. Based on the csi 300 index data, this paper selects the simplified artificial neural network model (BPNN), principal component analysis and neural network combination model (PCA-BPNN), genetic algorithm and neural network combination model (GA-BPNN), and improved genetic algorithm and neural network combination model (IGA-BPNN) for empirical analysis. The final results show that the improved IGA-BPNN model algorithm proposed in this paper can not only guarantee the prediction accuracy, but also control the number of influence variables.