Sentiment Analysis Based Stock Market Prediction Using Optimization Algorithm and Machine Learning

Abstract

Stock market price prediction is one of the most significant research areas in financial fields which influenced by political features, social factors as well as economic factors. In previously, several stock market prediction frameworks were developed, but that are having several issues. Therefore, this article presents a novel Long Short Terms Memory-Particle Swarm Whale Intelligence (LSTM-PSWI) algorithm were developed for stock market prediction using twitter data. Moreover, the twitter tweets are short and its having so many flaws, grammatical errors, and so data pre-processing crucial for to clean the unwanted data and convert it into the meaningful data. Hereafter, LSTM network is enabled to analyse the collected tweets into sentiment polarity such as positive and negative correlated with stock price.  Finally, whale and swarm optimization fitness were applied to the LSTM hidden state to classify the sentiment polarity. This model was implemented on the Python platform, and results were compared with traditional ML models. This LSTM-PSWI provides better performance for enhancing predictive analytics in stock market data through improved accuracy and robust analysis of large twitter text data. the attained outcomes are 11.71 RMSE, 9.2 % MAE and 10.4%  % MAPE