Financial Forecasting of Stock Market Using Sentiment Analysis and Data Analytics

Abstract

AbstractThe stock market prediction is the focus of many research works. State of art methodologies for stock prediction uses historical stock indices. News headlines, social media mentionings, and official reports influence stock market movements remarkably. The aim of this paper is to combine traditional data analytics methodologies and sentiment analysis to forecast the stock market trends. Along with machine learning classifiers we move one step forward and propose a system using deep learning methodologies and correlation of both the textual and numerical data analysis. We present the research done on prediction on stock trends using natural language processing. We further enhance the predictive model by integrating a sentiment analysis module on textual data to correlate the public sentiment of stock prices with the market trends. The experiments performed on real-world datasets conclude that Support Vector Machine (SVM), Random Forest Classifier, and Decision Tree Classifier performed well with more than 90% accuracy. For deep learning models, LSTM showed the highest accuracy (92%) followed by bidirectional RNN, deep CNN, shallow RNN neural networks. Our analysis shows that deep learning can be applied efficiently for stock market sentiment analysis, and the LSTM model is proven to be best performing on the textual data under study.KeywordsStock predictionClassificationDeep learningConvolutional neural networkRecurrent neural networkLong-short term memory