Stock Market Prediction via Twitter Sentiment Analysis using BERT

Abstract

The BERT and transformers models are utilised to analyse sentiment in this paper. The scope of this study is to decrease the risk of investing in the stock market using twitter sentiment analysis and to aid stockholders in making prudent investments that provide good returns. Then, after pre-processing tweet data, BERT and Word2Vec models were trained to get the expected outcomes. The attribute of this paper is to create a stock prediction method that will help stock market investors and inspire newcomers to the industry. The tweets and essential datasets are acquired via Twitter API and other sources because this paper is now focusing on stock price prediction from Twitter data. For stock prediction, the training dataset is transmitted to a pre-trained BERT model. Finally, the authors may use a combination of BERT and NLP to forecast if a particular stock will perform well in the future.