摘要
相关向量机(RVM)是一种基于贝叶斯模型的监督机器学习算法,可用于处理回归以及分类问题.与支持向量机(SVM)相比,相关向量机的优点在于其输出结果是一种概率模型,其相关向量的个数远远小于支持向量的个数,并且测试时间短.总结了相关向量机的基本原理及主要应用领域,详细阐述了相关向量机的模型结构以及分类方法,重点介绍了在高光谱图像分类中的应用.并针对RVM算法在高光谱图像分类中的不足,给出了多种改进方案,并作以比较.希望对研究者今后的研究有所启发,以促进该领域的发展.
The relevance vector machine （RVM） is a machine learning algorithm which is based on supervision of a Bayesian model. It can be used to deal with regression and classification problems. Compared with the support vector machine （SVM） , the relevance vector machine has the advantage that its output is a probability model and the number of relevance vectors is far fewer than the number of support vectors. In this paper, the application was sum- marized with a relevance vector machine and the classification of a hyperspectral image with RVM was introduced; the RVM model and the method of classification were also explained. In light of the disadvantage of classification, some improved methods were summarized. Various methods were generalized and analyzed while attempting to find breakthroughs and promote further research.
出处
《智能系统学报》
北大核心
2012年第4期294-301,共8页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金资助项目(61077079)
教育部博士点基金资助项目(20102304110013)
关键词
相关向量机
改进型相关向量机
高光谱图像
分类算法
relevance vector machine
improved relevance vector machine
hyperspectral image
classification algorithm
