融合情感因子的多因子选股模型构建与实证分析

Construction and Empirical Analysis of Multi-Factor Stock Selection Model Based on Affective Factor

摘要: 新型冠状病毒肺炎疫情背景下，投资者对医疗行业股票关注度增加，本文通过将代表投资者的情感倾向的情感因子加入到股票因子库中，研究情感因子的融入是否会优化选股效果。首先，选用医疗行业产业链主要股票作为候选股票池，提取241个因子数据，运用mRMR特征筛选算法进行因子优化，按照是否加入情感因子的对比预测方式，利用Stacking方法将机器学习模型进行融合后，构建多因子选股模型。通过对比模型结果，本文证实加入情感因子的模型的预测准确率更高。

Abstract: Under the background of the new coronavirus pneumonia epidemic, investors’ attention to medical industry stocks has increased, and this paper studies whether the integration of affec-tive factors will optimize the stock selection effect by adding affective factors representing investors’ emotional tendencies to the stock factor pool. Firstly, the main stocks of the medical industry chain are selected as the candidate stock pool, 241 factor data are extracted, the mRMR feature screening algorithm is used for factor optimization, and a multi-factor stock selection model is constructed after fusing the machine learning model according to the comparative prediction method of whether or not to add affective factors. By comparing the model results, this paper confirms that the model with affective factors has a higher prediction accuracy.