牛顿法在求解支持向量机中的应用

支持向量机作为一种重要的机器学习工具,近年来受到了广泛的关注,并得以迅速发展.但在处理大数据时,求解支持向量机对应的二次规划问题是非常棘手的,计算时间长,存储空间大.如何有效求解支持向量机是一个不可回避的研究课题.本文主要研究了如何利用牛顿法求解支持向量机和双生支持向量机,并提出了两个新算法.实验结果表明,所提算法是有效和高效的.

Application of Newton Method on Solving Support Vector Machines

Recently,as a kind of important machine learning tool,support vector machines have been attracted a widespread attention and obtained rapid development.But when dealing with large data,it is very difficult to solve the corresponding quadratic programming problems of support vector machines,which leads to take long computing time and have large storage space.It is an inevitable research subject how to effectively solve support vector machines.In this paper,we are devoted to study the problem of solving support vector machine and twin support vector machine by means of Newton method and provide two new algorithms.Experiment results indicate that our algorithms are effective and efficient.