Robust 1-Norm Soft Margin Smooth Support Vector Machine

Based on studies and experiments on the loss term of SVMs, we argue that 1-norm measurement is better than 2-norm measurement for outlier resistance. Thus, we modify the previous 2-norm soft margin smooth support vector machine (SSVM2) to propose a new 1-norm soft margin smooth support vector machine (SSVM1). Both SSVMs can be solved in primal form without a sophisticated optimization solver. We also propose a heuristic method for outlier filtering which costs little in training process and improves the ability of outlier resistance a lot. The experimental results show that SSVM1 with outlier filtering heuristic performs well not only on the clean, but also the polluted synthetic and benchmark UCI datasets.

• Type: Journal
• Year: 2010
• Where: IDEAL
• Authors: Li-Jen Chien, Yuh-Jye Lee, Zhi-Peng Kao, Chih-Cheng Chang
• Keywords: IDEAL 2010, Intelligent Agents, Margin Smooth Support, Outlier, Outlier Resistance