Prediction Of Stock Trading Point Of PLR-SVR Based On Box Theory

With the improvement of our national income,the per capita disposable income also increases,and the people urgently seek reliable investment way to realize the value of wealth.Stock investment because of its convenient,is being accepted by more and more people.China's stock market set up a short time,all aspects are not yet mature,the stock price movement has shown the characteristics of skyrocketing.The stock market crash so that investors can easily fall into a pessimistic mood,so that delay in entering the stock market.Stock market participation in investors is too low is not conducive to the healthy and orderly operation of the stock market,greatly weakening the stock market financing function.In order to help the investors reduce the investment losses,this paper puts forward the PLR-SVR stock trading system based on box theory,starting from the classical theory of stock trading.At present,China's stock market is in the process of change,Shanghai and Hong Kong through,Shenzhen and Hong Kong,the company forced to withdraw from the market,A shares into MSCIA and other major events,all show that China's stock market is gradually mature,gradually with international practice.Internationally,large-scale funds to hold blue-chip stocks,the future of China's blue chips will also occupy an increasingly important position in the stock market.In this paper,the most representative blue chip index-the composition of the SSE 50 index shares for the pool,from which 20 stocks as the object of this study.This paper presents the core idea of the trading system,namely:the formation of a stock price during the operation of the stock,the stock price has always fluctuated in the box up and down,when the stock price to run near the track to sell the stock,when the stock price to run near the next track to buy stock.This article uses the stock's 30-day maximum price(referred to as H30)as the upper track of the simulation trajectory,the use of stock within 30 days the lowest price(referred to as L30)as the next track of the simulation trajectory.When the closing price of the stock appears in the H30 neighborhood,the stock is sold and the stock is bought when the closing price of the stock appears in the L30 area.Because of the low number of data samples and the characteristics of high feature dimension,the support vector regression(SVR)is used to forecast H30 and L30.Considering that the transaction point of the trading system is essentially the turning point in the process of stock price operation,this paper uses piecewise linear representation(PLR)to denoise the stock price data.The advantage of PLR is that it divides the time series,preserves important timing turning points and maintains the basic trend of time series,so it is very suitable for the application scenario in this paper.Piecewise linear representation and support vector regression are used in stock price forecasting,which is one of the innovations of this paper.In the process of constructing the H30 prediction model and the L30 forecasting model by using the SVR algorithm,this paper combines the traditional technical index calculation concept with the stock trading volume to create the derivative index of the stock trading volume as the characteristic of the forecasting model,which is the second Innovation.In order to judge the profitability of the proposed stock trading system,this paper designs a series of comparative experiments to evaluate the historical data.In order to test whether the stock trading volume derivative index can effectively improve the forecasting accuracy,this paper will only include the SVR model with the stock price derivative characteristic and the SVR model which contains the derivative characteristic of the stock transaction.The test results show that the stock transaction volume The prediction accuracy of the posterior feature model is improved.In order to verify the profitability of the PLR-SVR stock trading system proposed in this paper,the PLR-SVR stock trading system is compared with the SVR stock trading system.The experimental results show that the threshold for segmenting the time series is critical.When selecting the appropriate threshold,PLR-SVR stock trading system profitability significantly more than SVR stock trading system.In order to verify the effectiveness of the proposed strategy,this paper chooses the classic buy and hold(BHS)strategy trading stock in stock trading.The experimental results show that the proposed trading strategy is better than BHS,whether it is PLR-SVR stock trading system or SVR stock trading system,its profitability are more than BHS more than doubled.