Predicting Sector Indices Using Machine Learning: A Comprehensive Analysis Incorporating Macroeconomic Indicators

The influence of macroeconomic indicators on stock market movement continues to attract attention, particularly in emerging markets, especially countries such as India. The objective of this study is to predict NIFTY 50 sectoral indices namely Automobile, Energy, Banking, IT, FMCG, Metal, and Realty using macroeconomic indicators such as exports, imports, exchange rates, GDP, interest discount rates, and inflation rates. Traditional regression models like Lasso, Linear Regression, Elastic Net, and advanced machine learning models like XGBoost, Decision Tree, and Random Forest Regressor, Support Vector Regressor have been used for the prediction. The experiments show that the Random Forest Regressor came out with the best performance, providing a lower Mean Squared Error (MSE) and the highest R-squared values in almost all the sectors, except the Realty sector. It proves that advanced machine learning techniques go beyond ordinary regression methods in capturing nonlinear relationships. This study highlights the importance of macroeconomic factors as drivers of sector-specific stock performances in emerging markets and the potential of using machine learning models to understand them.