Investor Sentiment Analysis Based on Comment Text for Stock Price Prediction

Abstract

The sentiment of stock investors is affected by many factors, and the trend of stock price is reflected by the decision behavior of stock trades. Previous tendency can be seen in technical index, and people's sentiment towards the stock is closely related to the text of commentary on the stock. This paper adopts sentiment dictionary to classify the same stock comment text, and modify the corpus of SnowNLP to train a model that conforms to Chinese stock comment. Based on the stock technical index and the established sentiment index to achieve the stock price prediction, so as to achieve a better analysis and prediction of the stock price trend. According to the analysis, the established investor sentiment index has a good effect on stock prediction. To be specific, the predictive regression score of the investor sentiment index is significantly improved compared with those of purely technical indicators, and MAE, MSE and other error indicators are obviously decreased. These results shed light on guiding further exploration of stock forecast with investor sentiment.