Application Of Support Vector Machine Based On Improved Cuckoo Algorithm In Stock Forecasting

In recent years,with the rapid development of China's economy,people have become increasingly interested in stock exchange.We expect to guide stock exchange through the analysis of all kinds of information in the stock market,so as to achieve wealth growth.However,the stock price is influenced by many factors,such as nonlinear,time-varying and highly unstable.It is a difficult problem for investors to choose rising stock from a large number of stock data.Firstly,the base of the analysis is the moving average line,the index smoothing difference average line and other technical indexes,and the stock data is predicted by establishing a better support vector machine model.The main work includes: improving the step size factor and the discovery probability of the cuckoo algorithm,and solving the effect of the search results of the cuckoo search algorithm by the step factor and the discovery probability.Then the improved cuckoo search algorithm is used to optimize the parameters and penalty factors of support vector machines,and a better support vector machine model is established to improve the quality of the prediction model and the accuracy of the model prediction.Secondly,in order to better analyze and verify the effectiveness and practicability of the proposed algorithm,this paper applies the improved algorithm to the prediction of the closing price and fluctuation of the stock data of the Pufa Bank,Shanghai airport,Huaxia Bank and Baiyun Airport,and compares the algorithm with the prediction results of other algorithms.The experimental results show that the support vector machine(ICS_SVM)based on the improved cuckoo algorithm is compared with the traditional support vector machine,the ICS_SVM consistency index(IA)and the accuracy rate are higher than the SVM based on the genetic algorithm and particle swarm optimization,while the mean square error(MSE)and the mean absolute error(MAE)are smaller,The mean deviation error(MBE)of CS_SVM tends to zero.In order to find a stable income point,the results of the forecast are further analyzed.The analysis shows that the higher the forecast increase results,the higher the accuracy.In view of this conclusion,this paper also carries out a classification experiment on stock data.The results show that when the classification results are positive,the approximate rate of stock volatility is rising,which provides a reference for investors in the stock investment.Finally,a stock prediction system based on Android is implemented in this paper.The whole system includes the server side and the Android end,in which the server realizes the acquisition and processing of the stock data,and provides the interface with the Android side,and the Android end realizes the data display.It provides five functions: stock market information,stock data visualization,stock forecast,financial news and user management.It can provide convenience for the analysis of ordinary investors.