A Review on Ensemble Learning Methods: Machine Learning Approach
Dipanshu Mishra, Shrikant Mani Tripathi, Akash Chaurasia, Pawan Kumar Chaurasia
Abstract
Unbalanced datasets make it difficult for machine learning to forecast the right class, but the ensemble approach offers a state-of-the-art workaround.Instead of basing predictions on a single model, the ensemble technique combines the pre-dictions of several models to forecast the proper class.This paper's goal is to examine the conventional ensemble approaches, such as bagging, boosting, and stacking generalization, and how they might be used to address the present problems posed by unbalanced datasets.To address these approaches' shortcomings, such as limited diversity in bagging, and overfitting in boosting, Researchers also discuss and compare other modifications of these classic techniques.From a variety of recently released works, Researchers highlight several new areas for ensemble classifier research.The study reviews several earlier theoretical research to offer a fuller understanding of the ensembles themselves.