The public sentiment analysis within big data distributed system for stock market prediction– a case study on colombo stock exchange

Abstract

Stock price prediction plays an important role on the journey of investors on the stock market. The prices of the company stocks on the market are performed by different deliverables. Social media data sets, news sites, feedback and reviews are some kind of online tools that can affect the stock market. It is often worth using this context to predict the performance of market shares. We take the advantage of Sentiment analysis on Market related announcement and respective public opinions for stock market trend predictions for more accurate recommendations. Sentiment Analysis is a machine learning program for extracting opinions from a text section that is designed to support any product, company, individual or other entity (positive, negatively, neutral). In this research calculations and data processing were performed within machine learning approach with use of Spark model on Google cloud platform. Among most of the stock prediction researches, only few researchers have done their researches on sentiment analysis within big data distributed environment. Logistic Regression and Naïve Bayes perform well in sentiment classification. Main finding of this research is that public opinion significantly influences the fluctuations of market forces and economic factors such as monetarism, government reforms, unforeseen pandemics, interest rates, public trust, and faith in bond market trust. The detection of the feelings pattern will enhance the market prediction as it ensures the consistency of decision.