Stock Trend Prediction for Energy Sector Stocks Based on Investor Sentiment and Deep Learning

The energy industry has always been one of the focuses of attention for investors and the media, among others, due to its special characteristics. Energy companies can be sensitively driven by a variety of factors such as policy changes, international situation and investor sentiment, etc. Traditional forecasting models are unable to make accurate predictions based on factors such as industry characteristics, and how to effectively capture these complex factors as well as more accurately predict stock prices in this industry based on these factors has become one of the important topics in financial research. In this study, we use historical trading data of energy stocks, take advantage of deep learning to process time series data and capture the dependencies between features, construct a stock prediction model with energy industry characteristics, and experimentally evaluate the model. The experimental results show that the prediction results are superior with the addition of industry characteristic indicators as well as indicators generated from texts related to the industry, which suggests that compared to generalized prediction models, models based on sentiment text analysis and deep learning are able to effectively identify and capture industry characteristics and make more accurate predictions.