Abstract: Some introductions will be made about common methods of SVM (Support Vector Machines) classification, some solution and improve also be discussed. Study a method of SVM to solve large-scale data through general SVM. Through theory’s analysis and data experiment prove that this method raise speed of calculus. Especially, when data set is large-scale this method will be more effective.

Key words: large-scale data set, speed of calculus, SVM, data classification algorithm

摘要： 对支持向量分类机中的一些基本方法做出详细的介绍，并进一步研究了方法的求解与改进。并通过对标准支持向量机的改造考虑了一种改进的方法，并进一步进行了相关的理论分析，通过数据实验验证了这种方法比传统的分类机在运算速度上有提高，特别是在处理较大规模的数据集时运算时间的效果更明显。

关键词: 大规模数据集, 运算速度, 支持向量机, 数据分类算法