Research on the Prediction Model of Stock Price Changes of China′s Science and Technology Innovation Board
In order to better predict the price changes of China′s sci-tech innovation board stocks,two machine learning al-gorithms,Random Forest and Support Vector Machine,were used to analyze the historical data of the Chinese stock market.318 Science and Technology Innovation Board stocks were selected as the sample dataset.Taking Huaxing Yuanchuang(688001.SH)as an example,the principles and algorithm processes of support vector machine and random forest were used to construct data samples,and the predictive effects were compared based on two approaches:yesterday′s closing price and the previous days′closing price.The results show that when predicting based on idea one,the accuracy of the random forest model is65.55%,and the accuracy of the support vector machine model is 70.59%.When predicting based on the second idea,the accuracy of the random forest model is43.70%,and the accuracy of the support vector machine model is62.18%;In terms of model selection,support vector machines are more suitable for predicting the level of stock market,and more vector machine models should be used to achieve prediction of China′s sci-tech innovation board stocks;In terms of indicator selection,the various indicators of the day are more informative than historical closing price data;And future results may not only be influenced by historical trends,but also by various indicators of the day.
NETL
NSTL
万方数据
维普
