Nifty Price Prediction from Nifty SGX using Machine Learning, Neural Networks and Sentiment Analysis

This paper investigates various Machine learning techniques such as Linear Regression, Decision Tree Regressor, Random Forest Regressor, and a neural network Multilayer Perceptron (MLP Regressor) to predict the opening price of the Nifty 50 index (on Indian National Stock Exchange (NSE)) based on the previous day’s closing price of the Singapore Exchange (SGX) Nifty index (which is traded on the Singapore Stock Exchange). The models use various parameters such as closing price of SGX Nifty, India volatility index (India VIX) and exchange value (of Singapore Dollar (SGD) to Indian Rupee (INR)) along with sentiment analysis based off thousands of tweets that included the hashtag "#Nifty50". The accuracy and errors for each model are calculated and compared. The impact of COVID-19 on the accuracy of model prediction is also analyzed. It is observed that the artificial neural network and linear regression offer the highest accuracy of 0.999, and the random forest regressor offers the lowest accuracy. COVID-19’s effect on the market is seen to impact the accuracy of model prediction, and the accuracy of prediction during the second wave is the lowest due to high market volatility averaged over a short period of time.