Constructing support vector machine ensemble with segmentation for imbalanced datasets |
| |
Authors: | Li Qian Yang Bing Li Yi Deng Naiyang Jing Ling |
| |
Affiliation: | 1.College of Science, China Agricultural University, Beijing, 100083, People’s Republic of China ;2.Department of Mathematics, Beijing University of Posts and Telecommunications, Beijing, 100876, People’s Republic of China ; |
| |
Abstract: | A novel method, namely ensemble support vector machine with segmentation (SeEn–SVM), for the classification of imbalanced datasets is proposed in this paper. In particular, vector quantization algorithm is used to segment the majority class and hence generates some small datasets that are of less imbalance than original one, and two different weighted functions are proposed to integrate all the results of basic classifiers. The goal of the SeEn–SVM algorithm is to improve the prediction accuracy of the minority class, which is more interesting for people. The SeEn–SVM is applied to six UCI datasets, and the results confirmed its better performance than previously proposed methods for imbalance problem. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|