摘要
近年来,计算机技术在金融投资决策中的应用越发深入,机构投资决策效率及准确性也因此显著提高,量化投资在国外已经是金融投资领域不可或缺的一股力量,基于机器学习的量化投资决策方法具有可观的应用潜力。围绕中证500指数,使用因子库中财务、量价、衍生等134个因子,通过PCA降维得到30个主成分因子,结合Adaboost模型进行一篮子个股买卖的预测。实证表明,相比纯Adaboost策略或指数表现,PCA_Adaboost具有更稳定的超额收益,因此,PCA_Adaboost投资策略具有广泛的应用前景。
In recent years,the application of computer technology in financial investment decision-making is more and more in-depth,and the efficiency and accuracy of institutional investment decision-making are also significantly improved.Quantitative investment has become an indispensable force in the field of financial investment abroad,and the quantitative investment decision-making method based on machine learning has considerable application potential.Based on the CSI 500 index,this paper uses 134 factors in the factor bank,such as finance,volume price,derivative,etc.,and obtains 30 principal component factors through PCA dimension reduction.Combined with AdaBoost model,it forecasts the trading of a basket of stocks.The empirical results show that compared with AdaBoost strategy or index performance,PCA_AdaBoost has a more stable excess return,so PCA_AdaBoost investment strategy has a wide application prospect.
作者
叶殷如
YE Yin-ru(Shenzhen Qianhai United Fortune Fund Management Co.,Ltd.,Shenzhen 518000,China)
出处
《佛山科学技术学院学报(自然科学版)》
CAS
2021年第4期76-80,共5页
Journal of Foshan University(Natural Science Edition)