An improved support vector machine with soft decision-making boundary

This paper proposes an improved support vector machine (SVM) classifier by introducing a soft decision-making boundary for solving real-world classification problem. The soft decision-making boundary contains two parameters describing the offset and the shape, which are estimated automatically from the distribution of training samples around the boundary via a distribution of belief degree in the decision value domain. The SVM with soft decision-making boundary increases classification accuracy by reducing the effects of data unbalance and noises in the real-world data. Simulation results show the effectiveness of the proposed approach.