中文摘要：

加权线性支持向量分类机是数据挖掘的新方法．它对应于一个优化问题．针对加权线性支持向量分类机优化问题建立了数据扰动分析理论方法．具体地针对加权线性支持向量分类机的原始问题建立了数据扰动分析基本定理，定理可以得到加权线性支持向量分类机问题的解及决策函数对数据参数的偏导数，同时可以定量分析输入数据的误差以及数据各种变化对其解以及决策函数值的定量影响，可以回答加权线性支持向量分类机问题的稳定性问题和灵敏度分析问题．

英文摘要：

Linear Support Vector Classifiers（SVC） is a new method for data mining. It is equal to a kind of optimal problem. The paper establishes SVC data perturbation analysis. In details, this paper establishes theory and method for linearly support vector classifier primal model. In accordance with primal model, basic theories of data perturbation analysis and the methords of partial derivative are formed for linearly support vector classifier dual problem. The partial derivative calculation of solution and decision function on data parameter is formed. Data perturbation analysis methods can be applied, with the partial derivative, to the analysis of the quantitative influence of data error on solution and decision function value. It can deal with the stabilitv and sensitivity analysis of linear SVC.