摘要
KNN(K-Nearest Neighbor)算法和贝叶斯网络分类算法(Bayesian Network,BN)都是目前应用非常广泛的分类算法。本文首先分析了KNN和BN的分类特点,然后在保留了两个算法在分类问题中优点的基础上,提出了基于贝叶斯网络结构学习的KNN算法(BN-KNN)。实验结果表明,BN-KNN算法能够有效地提高分类的正确率。
K-Nearest Neighbor algorithm(KNN) and Bayesian network classification algorithrn(BN) are currently widely used classification algorithms. At first, this paper analyzes the KNN and BN classified features, and then retains the merits of two classification algorithm. At last, a KNN algorithm based on learning the Bayesian network structure (BN-KNN) is presented. Experimental results show that BN-KNN algorithm can be used to improve the classification accuracy.
出处
《计算机科学》
CSCD
北大核心
2007年第12期184-186,237,共4页
Computer Science
基金
国家自然基金项目(60472017
30670699)资助课题
关键词
贝叶斯网络
K-近邻算法
距离加权
Bayesian network, K-nearest neighbor algorithm, Distance-weighted