摘要
人工智能在21世纪以来发展迅速,它与各个领域的结合发展促进了各个领域的飞快发展。本文重点研究人工智能在中国金融市场的量化应用,通过引入20个涵盖了价值、技术、动量、情绪反转等指标和8个机器学习算法对沪深两市股票收益率进行预测。从各个指标对模型的贡献程度来看,本文发现动量、反转和技术指标对股票未来收益率的影响程度最高。随后,本文按照这些股票的预测收益率进行排序并形成了交易策略。通过比较各个模型的结果发现预测收益率形成的交易策略在中国市场能获得显著的超额收益,且深度神经网络的预测效果最佳,正则化的线性机器学习模型次之。通过机器学习深度挖掘各个因子指标对中国股市的影响,为政策的制定者提供一定的借鉴意义,同时也能更好地理解中国市场交易中的非理性因素。
Artificial intelligence has developed rapidly since the 21st century,and its combination with the development of various fields promotes the rapid development of various fields.This paper introduces 20 indicators covering value,technology,momentum,sentiment,reversal and 8 machine learning algorithms to predict the returns of stocks in Shanghai and Shenzhen.From the perspective of the contribution of each index to the models,this paper finds that technical index and reversal index have the highest impact on the future stock return.Then,this paper sortes these stocks according to their predicted returns and forms a trading strategy.By comparing the results of various models,it is found that the trading strategy formed by predicting the return rate can obtain significant excess returns in the Chinese market,and the prediction effect of deep neural network is the best,followed by the regularized linear machine learning model.Through reasonable prediction methods,investors can avoid their own risks,meanwhile,it can provide certain reference significance for policy makers and better understand the irrational factors in Chinese market transactions.
作者
方毅
陈煜之
卫剑
Fang Yi;Chen Yuzhi;Wei Jian(School of Business and Management,Jilin University,Changchun 130000,China;Quantitative Economy Research Center,Jilin University,Changchun 130000,China;Guizhou University of Finance and Economics,Guiyang 550000,China)
出处
《工业技术经济》
北大核心
2022年第8期83-91,共9页
Journal of Industrial Technological Economics
基金
国家自然科学基金面上项目“基于随机占优的高阶偏好投资组合构建”(项目编号:71871104)。
关键词
机器学习
资产定价
股票收益
市场异象
预测
交易策略
machine learning
asset pricing
stock return
market anomalies
prediction
trade strategies