基于股吧情感本体与传播的股市短线预测

为研究股吧文本中蕴含的股民情感对股市短线走势的影响，使用情感倾向点互信息算法构建面向股市的情感词典，分别考虑情感本体与情感来源传播建立面向股市的加权情感倾向得分模型，并进一步建立ARMA-GARCHX模型以分析股民情感与股市短线走势之间的关系.研究显示：面向股市的情感词典相比于传统情感词典具有更强的适应性，使用加权情感倾向得分模型建立的预测模型对股市走向的预测相比于传统方法具有更优良的预测效果.

Research on Stock Market Short-Term Prediction Based on Stock BBS Emotional Ontology and Dissemination

In order to research the impact of stockholder sentiment contained in the stock BBS text on the short-term trend of the stock market, the sentiment dictionary facing the stock market is constructed by using the semantic orientation pointwise mutual information. The weighted emotion tendency score model facing the stock market is established considering emotional ontology and source dissemination respectively, and the ARMA-GARCHX model is further established to analyze the relationship between the stockholder sentiment and the short-term trend of stock market. The research shows that the sentiment dictionary facing the stock market has stronger adaptability than the traditional sentiment dictionary, and the prediction model built by the weighted emotion tendency score model has better prediction effect than the traditional method.