Feature selection for linear support vector machines

Feature selection is attracted much interest from researchers in many fields such as pattern recognition and data mining. In this paper, a novel algorithm for feature selection is developed. The proposed algorithm uses the standard linear SVM algorithm and is performed in an iterative way. Feature selection is carried out by assigning weights to features. Experimental results on UCI data set and face images confirm the feasibility and validation of the proposed method.

• Type: Conference
• Year: 2006
• Where: ICPR
• Authors: Zhizheng Liang, Tuo Zhao
• Keywords: Computer Vision, Feature Selection, ICPR 2006, Linear Svm Algorithm, UCI Data