RVM的多类分类算法及其在钙化点形态分类中的应用
首发时间:2008-03-17
摘要:关联向量机(RVM)是近年来提出的机器学习算法,它基于贝叶斯学习和核函数方法,可用于分类与回归中。本文在RVM二类分类的基础上,提出了将RVM扩展为多类分类的算法,从而不必显式地去构造多个二类分类器。钙化是乳腺X光图像中的乳腺癌的早期征象之一,钙化点的形态与其良恶性之间有重要的关系。本文探讨了如何将RVM多类分类算法运用于钙化点的形态分类,将做了相关实验。
For information in English, please click here
Multi-class RVM Algorithm and its Use in Shape Classification of Calcifications
Abstract:Relevant vector machine is a new machine learning algorithm. It based on Bayesian learning and kernel function method and can be used in classification and regress. The paper expands the two-class RVM classification algorithm to multi-class RVM classification algorithm. There is no need to construct multiple two-class classifiers. Calcification is one of the primary symptoms of early breast cancer and its shape is related to weather it is benign and malignant. The paper discussed how to use the multi-class RVM algorithm on the shape classification of calcification and had several experiments.
Keywords: relevant vector machine multi-class classification calcifications shape classification
基金:
论文图表:
引用
No.1937619274212057****
同行评议
共计0人参与
勘误表
RVM的多类分类算法及其在钙化点形态分类中的应用
评论
全部评论0/1000