A Dynamic Cost Sensitive Support Vector Machine
Xiao Lin Chen, Yan Jiang, Min Jie Chen, Yong Hong Yu, Hong Nie, Min Li · Advanced materials research · 2012
A lot of cost-sensitive support machine vector methods are used to handle the imbalanced datasets, but the obtained results are not as perfect as expectation. A promising method is proposed in this paper, named ADC-SVM, which uses genetic algorithm to dynamically search the optimal misclassification cost to build a cost sensitive support machine. We empirically evaluate ADC-SVM with SVM and Cost-sensitive SVM over 8 realistic imbalanced bi-class datasets from UCI. The experimental results show that ADC-SVM outperforms the other two methods over all the imbalanced datasets.