Comparative Study on the Performance of the Bagging Algorithm in the Breast Cancer Dataset

Abstract

Breast cancer remains a predominant health concern globally. Early detection, powered by advancements in medical imaging and computational methods, plays a vital role in enhancing survival rates. This research delved into the application and performance of the Bagging algorithm on a Breast Cancer dataset that underwent image segmentation using the Canny method and feature extraction through Hu-Moments. The Bagging algorithm demonstrated moderately consistent performance across a 5-fold cross-validation, with average metrics of 56.9% accuracy, 58.3% precision, 57.7% recall, and 56.6% F-measure. While the results showcased the potential of the Bagging algorithm in classifying breast cancer data, there remains an avenue for further optimization and exploration of other ensemble or deep learning techniques. The findings contribute to the broader domain of machine learning in medical imaging and offer insights for future research directions and clinical diagnostic tool development.