摘要
阐述了核方法的基本原理与研究动机,分析了特征空间的性质,介绍了常见的核方法,给出了构建新核方法的步骤及需要注意的问题,指出了核方法值得关注的研究方向,展示了其在多用户检测中的应用情况,以其对核方法研究领域有较全面的把握。
The major characteristics of the feature space and present alternative methods and corresponding algorithms were analyzed. The steps to construct a novel kernel method and the future research issues were given. Finally the applications to multi-user detection using KM were explored. It is expected to understand KM comprehensively.
出处
《通信学报》
EI
CSCD
北大核心
2005年第7期96-108,共13页
Journal on Communications
基金
国家自然科学基金资助项目(40274019)
关键词
核方法
支持向量机
机器学习
再生核希尔伯特空间
多用户检测
kernel method
support vector machine
machines learning
reproducing kernel Hilbert space
multi-user detection