Construction of Quantization Strategy Based on Random Forest and XGBoost
Mingli Zhu · 2020
Quantitative investment refers to an important decision that uses a combination of digital and computer algorithms to help investors make decisions. It has a history of 30 years. As computer algorithms become more and more intelligent, a number of investors in modern times are making quantitative investment with the aid of artificial intelligence. This article discusses methods for predicting the stock movements based on the machine learning. The prediction results obtained from two models of random forest and XGBoost are compared. The historical closing data of the Shanghai and Shenzhen 300 shares at the end of the current month, which acts as the source of data for the prediction model, was calculated based on the EMA, RSI and ROC indicators to obtain the relatively weak indicators, the receiver's working characteristic curve and the exponential moving average. According to the random forest and XGBoost accuracy rate, the accuracy rate and the F score are compared to obtain the experimental results. The accuracy of XGBoost is 78%, and the accuracy of random forest is 72%. Experimental results show that, compared to Random Forest, the yield and sharp ratio of XGBoost are 6% higher. The model employed in this paper incorporates the disadvantages of the linear model, and with the model, the stock profit is verified and the stock risk is predicted, which provides a reference for the investor in their decision-making.