Abstract

This study combines sentiment analysis, autoregressive integrated moving average (ARIMA) models, recurrent neural network (RNN) and long short-term memory (LSTM) to provide a thorough method of stock price prediction. In order to improve the precision and resilience of stock price projections, the study attempts to capitalize on the advantages of each approach. The first step of the investigation is gathering important textual material and historical stock price data from financial news, social media, and other sources. A strong foundation for capturing both short-term dependence and long-term trends in stock prices is provided by the combination of LSTM and ARIMA models. Utilizing natural language processing (NLP) methods, the sentiment analysis component draws insightful conclusions from textual data and incorporates market sentiment as a critical component of the prediction model.