A linear multivariate binary decision tree classifier based on K-means splitting

Fei Wang, Quan Wang, Feiping Nie, Zhongheng Li, Weizhong Yu, Fuji Ren

Abstract: Highlights•Introducing the unsupervised K-means clustering into supervised tree classification framework improves the classification performance.•The proposed non-split condition enables the tree model to accommodate the class imbalance cases more flexibly.•The proposed tree model is efficient during training and classification.