A Hybrid Approach Based on Decision Trees and Clustering for Breast Cancer Classification

This paper proposes a hybrid diagnosis approach of breast cancer based on decision trees and clustering. Our proposed approach does not only assume distinguishing malignant from benign cases, but also makes a refined treatment of these latter. Experimental study on Wisconsin Breast Cancer Database provides a thorough analysis of the induced results and shows that we can enhance the classification results by distinguishing different types of Breast Cancer using a clustering technique.