Multiclass SVM with ramp loss for imbalanced data classification

Class imbalance is a common problem encountered in applying machine learning tools to real-world data. It causes most classifiers to perform sub-optimally and yield very poor performance when a dataset is highly imbalance. In this paper, we study a new method of formulating a multiclass SVM problem for imbalanced dataset to improve the classification performance. The proposed method applies cost-sensitive approach and ramp loss function to the Crammer & Singer multiclass SVM formulation. Experimental results on multiple UCI datasets show that the proposed solution can effectively cure the problem when the datasets are noisy and highly imbalanced.