A Stock Price Prediction Model Based on Investor Sentiment and Optimized Deep Learning

Accurate prediction of stock prices can reduce investment risks and increase returns. This paper combines the multi-source data affecting applies sentiment analysis, swarm intelligence algorithm, deep learning to build MS-SSA-LSTM model. Firstly, we crawl East Money forum posts information establish unique dictionary calculate index. Then, Sparrow Search Algorithm (SSA) optimizes Long Short-Term Memory network (LSTM) hyperparameters. Finally, index fundamental trading are integrated, LSTM is used forecast in future. Experiments demonstrate that model outperforms others has high universal applicability. Compared with standard LSTM, $R^{2}$ improved by 10.74% on average. We found that: 1) Adding enhance model’s predictive performance. 2) The LSTM’s hyperparameters optimized using SSA, which objectively explains parameter settings improves effect. 3) volatility China’s financial market more suitable for short-term prediction.