Hybrid Dataset Trained LSTM Model for Forecasting Stock Market Trends: Analyzing Sentiment of Tweets and Nifty 50 Index Values
摘要
This research paper proposes a hybrid method that combines historical stock data with sentiment analysis of Twitter data to predict the stock market index. Twitter has become a popular platform for expressing public opinion and sentiment about various events, including the stock market. The study employs various steps such as preprocessing, hyperparameter tuning, and Long Short-Term Memory (LSTM) modeling to analyze the sentiment of tweets and forecast Nifty 50 index values. The proposed approach provides a promising direction for predicting stock market indices, particularly based on short-term prediction using social media data. The hybrid dataset combining historical stock index values with Twitter text data leads to improved accuracy in forecasting stock market indices and trends.