机构地区: 辽宁师范大学数学学院
出 处: 《系统工程学报》 2005年第1期109-112,共4页
摘 要: 支持向量机是一种基于统计学习理论的新颖的机器学习方法,由于其出色的学习性能,该技术已成为当前国际机器学习界的研究热点.这种方法已广泛用于解决分类和回归问题.在回归中,标准的支持向量回归算法在采用ε-不敏感损失函数时引入两个参数.为了减小学习复杂性,给出一种单参数约束下的支持向量回归算法,该算法能够减少支持向量的数量,提高程序的运行速度.最后,以一个混沌时间序列预测为例,所给方法同标准支持向量回归算法进行了比较,运行速度明显提高. Support vector machines(SVM) are a kind of novel machine learning methods, based on statistical learning theory, which have become the hotspot of machine learning because of their excellent learning performance. The method of support vector machines has been developed for solving classification and regression problems. In the case of regression, standard support vector regression algorithm usually introduces two parameters in using ε-insensitive loss function. In order to reduce complexity, a kind of support vector regression algorithm is proposed by introducing a single parameter. This method can reduce the number of support vectors, and improve the speed of running. Finally, the proposed method is compared with the standard support vector regression algorithm by predicting a chaotic time series, and its speed of running is improved obviously.
领 域: [自动化与计算机技术] [自动化与计算机技术]