Adjusting Method for Imbalanced Support Vector Machines
不平衡支持向量机的调整方法

An adjusting method for the offset of separating hyperplane of binary-classification imbalanced data was proposed.Firstly,the two class of samples were trained preliminarily and a normal vector of the separating hyperplane was generated.Secondly one dimensional data were obtained by projecting the high dimensional sample data onto the normal vector.Finally a new algorithm was proposed by using the information provided by the estimation of the non-parameter density for the projection data.Thus the offset of ...