Performance Evaluation of Improved Adaboost Framework in Randomized Phases Through Stumps
D Sudharson, Shameem A. Fathima, Pallavi Kailas, K S Vaishnavi, Subhashree Darshana, A Bhuvaneshwaran · 2021 International Conference on Advancements in Electrical, Electronics, Communication, Computing and Automation (ICAECA) · 2021
AdaBoost algorithm is being a boosting framework in Ensemble Method that is extensively used in the streams of machine learning. In supervised learning, boosting helps to reduce bias and variance. Adaptive Boosting works in a way such that the weights are re-assigned to every instance with higher weights to faulty classified instances. Classification problems with numerous classes in an imbalanced dataset presents a predominant challenge than a binary classification. In the conventional machine learning algorithms, the skewed distribution makes numerous impacts on the effectiveness when the prediction is restricted to the examples of the minority class. Multiclass Classification refers to the classification process involving more than two classes. When there is imbalanced data that mostly relates to a classification problem in which the classes are represented equally. Hence, an improved AdaBoost algorithm can solve the above issues accordingly.