基于熵权法和CART决策树的量化选股策略

Quantitative Stock Selection Strategy Based on Entropy Method and CART Decision Tree

本文以粤港澳大湾区的30支股票为研究样本,结合券商研报运用熵权法综合考虑多层面因素建立影响股票走势的特征指标体系并探究其对公司股票走势的影响。经检验,从短期来看:资本市场处于弱式,且并不总是有效的;从中长期来看:只有市场整体、估值因子对股票走势有显著负向影响,其他指标未证实与回报率有相关关系。根据特征指标体系建立基于CART决策树的选股模型,随机抽取两组股票进行模型测试与检验,为投资者提供更优的选股策略。考虑外部环境因素对股票走势的作用机制后引入基于新冠疫情的稳定能力因子完善特征指标体系,以此进一步优化选股模型并对决策结果进行测试与检验。

This paper uses 30 stocks in the Guangdong-Hong Kong-Macao Greater Bay Area as a research sample,combined with the research reports of securities companies, using the entropy method to comprehensively consider multiple factors to establish a characteristic index system that affects the stock trend and explore its impact on the company’s stock trend. After testing, from a short-term perspective: the capital market is weak and not always effective;from a mid-to-long term perspective: only the overall market and valuation factors have a significant negative impact on stock trends. A stock selection model based on the CART decision tree is established according to the characteristic index system, and two sets of stocks are randomly selected for model testing and verification, so as to provide investors with better stock selection strategies. Taking into account the mechanism of external environmental factors on stock trends, the stable capability factor based on the COVID-19 is introduced to improve the characteristic index system, so as to further optimize the stock selection model and test and verify the decision-making results.