Decision tree classifiers for mass classification

摘要

Mass detection from the mammogram is important for breast cancer diagnosis. This paper proposes the classification method for breast masses using the decision tree techniques. This paper presents the comparison result of 12 decision tree algorithms including ADTree, BFTree, DecisionStump, FT, C4.5, LADTree, LMT, NBTree, RandomForest, RandomTree, REPTree and CART. In comparison, four performance metrics were used. The aim of the study is to determine the best decision tree classifier for mass classification from BI-RADS features (mass shape, mass margin, assessment and subtlety). In the experimental studies, all these decision tree algorithms are applied on the UCI data set. Experimental results show that LADTree and LMT has a better performance than ADTree, BFTree, DecisionStump, FT, C4.5, NBTree, RandomForest, Random Tree, REPTree and CART.