关键词: 方向预测, 情绪指数, 评分加权

Abstract: The development of the financial market concerns the economic lifeline of a country. As the stock market is an important part of the financial market, the research on the stock returns has always been a hot topic academically. Financial news is often considered to contain rich information, and as one of the important factors affecting investors’ investment decisions, the emotion contained in financial news also has a certain impact on stock returns. Therefore, this paper constructs a sentiment lexicon of financial news, which is applicable in the financial investment field, and uses it to analyze the text of financial news. A new prediction model is established: to quantify the sentiment contained in financial news text as a sentiment index, and combine it with the time-varying density function to obtain a time-varying weighted density model. On this basis, scoring the model and using the model score as the weight to combine different prediction models to construct a scoring weighted model for stock returns prediction. The results show that the sentiment index of financial news can improve the prediction of the model, and the prediction effect of the scoring weighted model is double optimal in accuracy and scoring rules.

Key words: Direction Prediction, Sentiment Index, Scoring Weighted