TSVR Algorithm Based On GA Optimization And It's Application In Shanghai Stock Index Prediction

Since the first joint stock limited company was established in China,with the development of China's economy and the support of national policies,investment has become popular throughout the country.Stock investment has become a part of the national life,and the stock market has also become an important part of the national financial market.How to effectively reduce the risk of investment and expand the return on investment has become a hot topic for shareholders,which is also one of the hot topics for many scholars.Therefore,the effective prediction of stock trends can not only provide investors with investment basis,but also provide new algorithm reference for the field of stock prediction.In recent years,with the rapid development of machine learning and artificial intelligence,more and more shareholders begin to invest from the technical analysis level,which makes the domestic stock analysis technology develop rapidly.In this paper,a new algorithm,twin support vector regression(GA-TSVR)algorithm based on genetic algorithm optimization,is proposed in the field of stock prediction.It's main idea is: First,genetic algorithm is used to optimize the parameters of twin support vector regression algorithm to find the optimal penalty factor and kernel parameters;Second,we use R to capture the stock index data from Yahoo! Webpage,and calculate short-term technical indicators such as moving average(MA),exponential moving average(EMA),smooth similarity and difference moving average(MACD),relative strength index(RSI),rate of change(ROC),William index(W%R)and integrate them with the original stock data;Third,the results obtained by the genetic algorithm and the sorted data are substituted into the model for prediction and evaluation.In order to ensure the authenticity of prediction results,the results of twin support vector regression and support vector regression are compared.The results show that GA-TSVR model has certain advantages in prediction accuracy and prediction time.