基于R语言的网络舆情对股市影响研究  

Study on the impact of network public opinion on the stock market based on R-language

作　　者：朱昶胜[1] 孙欣 冯文芳[2] ZHU Chang-sheng;SUN Xin;FENG Wen-fang(School of Computer and Communication,Lanzhou Univ.of Tech,Lanzhou 730050,China;School of Economics and Management Lanzhou Univ.of Tech,Lanzhou 730050,China)

机构地区：[1]兰州理工大学计算机与通信学院,甘肃兰州730050 [2]兰州理工大学经济管理学院,甘肃兰州730050

出　　处：《兰州理工大学学报》2018年第4期103-108,共6页Journal of Lanzhou University of Technology

基　　金：兰州理工大学红柳杰出人才基金项目(J201304)的资助

摘　　要：以开源R语言为平台,东方财富网的股评为研究对象,结合中文文本挖掘技术和SVR支持向量回归模型.利用中文挖掘技术,对股评进行去噪声、分词、同义词合并、去停用词、TFIDF、文本向量化将非结构化文本数据转化为结构化的特征向量矩阵,与股票的收益率建立SVR回归模型,通过预测未来的股票收益率来预测股价的涨跌趋势.研究结果表明,预测股价涨跌趋势与实际趋势基本吻合,可以通过分析网络舆情来对股市未来发展趋势进行预测.The open source R language is taken as the platform, the Oriental Fortune Network is taken as the research object, and the Chinese text mining technique and support vector regression（SVR） model are incorporated to examine the stock market. By means of Chinese text mining technique, thedenoising,word segmenting, synonyms merging, waste word descarding, TFIDF, and text vectorization are carried out for the stock review text, the unstructuralized text data is transformed into a structuralized eigenvector ma trix, the SVR regression model is established along with the return rate of the stock, and the volatility trend of stock price is predicted by means of prediction of the future return rate of the stock. The investi gation result shows that the predicted volatility trend of the stock price will basically coincide with the ac

关 键 词：网络舆情 R语言 中文文本挖掘 SVR模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]