MARGIN MAXIMIZATION MODEL OF TEXT CLASSIFICATION BASED ON SUPPORT VECTOR MACHINES
Support Vector Machines (SVMs) are more suitable for text categorization than traditional machine learning methods by acknowledging various statistical characteristic of text learning task. By introducing the margin maximization principle in the statistical machine learning theory, the feature statistic matrix based on average document mapping (FSM-ADM) model, which partitions the set of features using weighted odds ratio, is proposed in the form of generalization capability estimationtheorem with rigorous proofs and solid experimental validation. The theoretical model has successfully discovered the unexplored capability of being classified in text classification.
Support vector machines text classification FSM-ADM margin maximization
PENG CHEN TAO WEN
Department of Computer Science & Technology, Neusoft Institute of Information, Dalian 116023, China
国际会议
2006 International Conference on Machine Learning and Cybernetics(IEEE第五届机器学习与控制论坛)
大连
英文
3514-3518
2006-08-13（万方平台首次上网日期，不代表论文的发表时间）