Abstract: In ordinary Support Vector Machine，for each elment（x_i，y_i），i=1，2，…，l of a training set，input of x_i has a definite class number y_i.While in practice，class of x_i is rarely fixed，x_i generally belongs to positive Class with probability of z_i⁺ or xi generally belongs to negative class with probability of z_i^-.In that light，this paper has modified the training set，In which yi is replaced by z_i⁺，z_i^- and classification optimal model and Uncertain Proximal Support Vector Machine are construct in turn.

Key words: uncertain, proximal, Support Vector Machine

摘要： 在一般支持向量机中，训练集的每个元素（x_i，y_i），i=1，2，…，l，对输入x_i有确定的类别标号y_i 。而在许多实际问题中，x_i的类别往往是不确定的，常常是以概率z_i⁺属于正类的概率z_i^-属于负类。针对这种实际情况，改造训练集，把原来的y_i，用z_i⁺，z_i^-代替，然后建立分类最优先模型，构建不确定中心支持向量机。

关键词: 不确定, 中心, 支持向量机