Using Financial News Headlines to Predict Stock Data and Conduct Sentiment Analysis

Abstract: This study offers a web application for stock prediction that uses machine learning algorithms to forecast market performance and analyse the sentiment of financial news. LSTM, lexicon-based analysis, and vector-based machine learning approaches are employed. for stock forecasting and sentiment analysis. The sentiment analysis system classifies news headlines as either positive or negative with an accuracy rate of 86%. The mean absolute error of the stock prediction model's stock performance estimation is 3.4 percent. The total accuracy of both models in predicting stock performance is 83%. The results indicate the system's potential for use in the stock market by offering crucial insights into machine learning algorithms for sentiment analysis using financial news headlines and stock data prediction