摘要
Boosting是在统计学习理论的基础上发展起来的一种新的集成机器学习方法，并在模式分类领域有了广泛的应用。该文首先分析了Boosting的原理并介绍了其经典算法AdaBoost方法，分别引入三种核函数（多项式核函数、径向基核函数、Sigmoid核函数）集成AdaBoost算法的弱分类器。然后将其应用于两个关于癌症论断的数据集中，通过实验验证了核函数作为弱分类器集成AdaBoost分类器的良好性能。
Boosting is a new method of ensemble machine learning developed from the theory of Statistical Learning Theory （STL）, and it has a great application on the pattern classification field. The theory of Boosting and the classical algorithm of AdaBoost are studied at first, then it is introduced three kinds of kernel function （Polynomial kernel, R.adial Basis Function, Sigmoid kernel function） integrated weak classifier for AdaBoost. Then it applied to two conclusions about cancer data set by experimental verification of a nuclear function as a weak classifier integrates the good performance of AdaBoost classifiers.
作者
李想
李涛
LI Xiang, LI Tao （College of Automation and Electrical Engineering, Lanzhou Jiaotong University, Lanzhou 730070, China）
出处
《电脑知识与技术》
2011年第10期6969-6970,6976,共3页
Computer Knowledge and Technology
