Neural Network Prediction Model Based On Bat Algorithm Research And Application

As an important representative in the economic field,the stock market has attracted extensive attention from all over the world.In addition,because stock prices are affected by many factors,stock prices are highly uncertain.Therefore,scientific prediction of stock prices can guide investors to make reasonable investment and minimize investment risks.Based on neural network,a new artificial intelligence Algorithm—Bat Algorithm(BA)is combined with BP neural network,and a neural network prediction model based on improved Bat Algorithm and the optimization model are studied.The results have certain theoretical significance and practical application value.The main research work of this paper is as follows:1.On the basis of summarizing the BP neural network and Bat Algorithm and the research status at home and abroad,and combining with the theory of modern Artificial Intelligence and the characteristics of stock price,it provides basic reference for the establishment of nonlinear time series data prediction model.2.In order to improve the convergence speed,optimization accuracy and initial solution quality of the traditional Bat Algorithm,Chaos Optimization Algorithm is introduced to optimize the Bat Algorithm,and then the neural network is trained.A neural network prediction model based on Chaos Bat Algorithm(CBA)is proposed.Considering that Chaos Bat Algorithm still has the disadvantages of single species and easy to fall into local optimal,hence Differential Evolution Algorithm is further introduced to train neural network by using Chaos Difference Evolution Bat Algorithm,and a neural network prediction model based on Chaos Difference Evolution Bat Algorithm(CDEBA)is proposed.3.Anshan iron and steel stock price(000898)data is taken as a research example,the neural network prediction model and improved model based on Bat Algorithm are used to analyze the prediction,and the feasibility and effectiveness of the improved prediction model are verified to provide a new idea for the research of stock price prediction method.