Trend Prediction And Quantitative Strategy Design Based On K-line Chart Machine Recognition

The K-line chart is an important research area for graphic analysis in technical analysis.Over the years,it has accumulated a series of analysis and application rules through the application summary of previous investors.Line graph analysis relies heavily on subjective factors such as the personal experience of the analyst.The same person's application of unified rules often results in different or even opposite results.Therefore,as a K-line diagram representing historical information,whether it really contains information that can predict future price changes has become a research direction worthy of attention.Therefore,this paper focuses on whether the k-line graph,which represents historical information,really contains information that can predict future price movements.With the extensive study of deep learning in recent years,the application of deep neural networks to stock market forecasting has also become a popular research direction.In existing research,time price series are used as the input form of the network.A study of line graphs combined with neural networks.Convolutional neural networks have unique advantages in picture classification and recognition.Therefore,in the research of this paper,the unique advantages of convolutional neural networks in picture recognition are compared with K-line chart,the most important quantity and price information recognition for stocks.Combining and applying the experimental ideas of the control variable method,a series of comparative models are set up to conduct research through conditional control,and try to verify from the side whether the Kline chart really contains information that can predict future price changes.In the end,the research in this paper found that the advantages of convolutional neural networks in image recognition were applied to the classification and recognition of K-line diagrams.This research method is feasible.In the model training in this paper,the 20-day K-line chart plus the volume indicator is used as the trading cycle.The 50*50-dimensional convolutional neural network training has the best prediction effect and the prediction accuracy reaches 86.4%.Subsequently,the convolutional neural network model constructed in this paper was applied to the actual operation of the quantitative trading strategy.Tests on the trading platform proved that the model is effective.In the one-year trading platform back test,short selling is not allowed The result of annualized return is 36.73%,and the maximum drawdown rate is 29.6%.Under the circumstances of allowing shorts,the annualized return was 61.27%,and the maximum drawdown rate was 20.37%.Therefore,the convolutional neural network model constructed in this paper is effective and can obtain certain benefits in the quantitative practical application of recognizing the prediction of the rise and fall of the K-Line chart.