Using Social Network Sentiment Analysis and Genetic Algorithm to Improve the Stock Prediction Accuracy of the Deep Learning-Based Approach

Abstract Traditionally, most investment tools used to predict stocks are based on quantitative variables, such as finance and capital flow. With the widespread impact of Internet, investors institutions designing strategies also referring online comments discussions. However, multiple information sources, along with uncertainties accompanying international political economic events recent pandemic, have left concerned about interpretation approaches that could aid decision-making. To this end, study proposes a method combines social media sentiment, genetic algorithm (GA), deep learning changes in stock prices. First, it employs hybrid (HGA) combined machine identify chip-based indicators closely related fluctuations prices then uses them input for long short-term memory (LSTM) establish prediction model. Next, five sentiment variables analyze PTT TSMC’s price performs grey relational analysis (GRA) fluctuations. The selected build LSTM improve efficiency analysis, applies Taguchi optimize hyper-parameters. results show proposed using HGA-screened an model can effectively accuracy