In view of inconsistent problems caused by that Synthetic Minority Over-sampling Technique (SMOTE) and Support Vector Machine (SVM) work in different space, this paper presents a kernel-based SMOTE approach to solve classification with imbalance data set by SVM. The method first preprocesses the data by oversampling the minority instances in the feature space, then the pre-images of the synthetic samples are found based on a distance relation between feature space and input space. Finally, these pre-images are appended to the original dataset to train a SVM. Experiments on real data set indicate that compared with SMOTE approach, the samples constructed by the proposed method have the higher quality. As a result, the effectiveness of classification by SVM on imbalance data set is improved. © 2009 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Zeng, Z. Q., & Gao, J. (2009). Improving SVM classification with imbalance data set. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5863 LNCS, pp. 389–398). https://doi.org/10.1007/978-3-642-10677-4_44
Mendeley helps you to discover research relevant for your work.