基于贝叶斯分类法的股票选择模型的研究  

Research on Stock Selection Model Based on Bayesian Classifier

作　　者：骆桦[1] 张喜梅[1] 

机构地区：[1]浙江理工大学理学院,杭州310018

出　　处：《浙江理工大学学报（自然科学版）》2015年第3期418-422,共5页Journal of Zhejiang Sci-Tech University(Natural Sciences)

基　　金：国家自然科学基金项目(11401532)

摘　　要：股票选择在证券投资中是一个重要问题,提出了一种基于朴素贝叶斯分类方法的股票选择模型。首先根据美股特斯拉的表现对沪深证券市场的能源股进行聚类分析,选取对股票投资价值影响显著的财务指标构造样本特征集,其次通过合理地选取贝叶斯分类器的参数对股票进行分类。经研究所选取股票的等权投资组合产生了44.6%的累计回报率,优于32.4%的基准回报率,表明朴素贝叶斯分类方法在股票选取方面具有较好的效果,值得投资者借鉴。Stock selection is an important problem in the securities investment. This paper proposes a stock selection model based on naive Bayesian classification method. Firstly, according to the performance of the US stocks Tesla, cluster analysis is made for energy stock on Shanghai and Shenzhen securities markets. Financial indexes which impose significant influenced on stock investment value are selected to construct the sample set. Secondly, we classify stocks by choosing the appropriate parameters of naive Bayesian classifier. After study, the equally weighted portfolio of stocks yields accumulative return rate of 44. 6G, which is better than the benchmark return rate of 32. 4%. The results show that naive Bayesian classification method has good effects on stock selection and investors should refer to it.

关 键 词：股票选择 投资决策 聚类分析 朴素贝叶斯分类方法 

分 类 号：F830.91[经济管理—金融学]