Ensemble Machine Learning Approach for Performance Evaluation of Mutual Funds (Index) with news-based Sentiment
TL;DR
The research showed that hybrid approaches to machine learning can substantially improve index fund predictions, offering insights for investors and financial experts in turbulent market environments.
Abstract
This study proposes a hybrid machine learning framework for predicting equity index fund performance based on ensemble regression. It aims to amalgamate the forecasting accuracy by combining the machine learning algorithms viz., SVR, KNN and Random Forest Regression under a Voting Regressor model. The hybrid algorithm combines the strength of individual models to reduce the prediction errors and improve the generalizations. The models were evaluated using metrics such as R2, MSE, RMSE and MAPE The study showed that the hybrid Voting Regressor had higher prediction accuracy and stability on several index funds than the individual models. The ensemble method yielded higher R² values and lower forecast errors, suggesting its capacity to capture complex market dynamics and reduce bias. The research showed that hybrid approaches to machine learning can substantially improve index fund predictions, offering insights for investors and financial experts in turbulent market environments.