一种k-最临近算法的设计、实现与评测
esign, Implementation and Evaluation of a k-Nearest Algorithm
本文首先简要的介绍了k-最临近（也称为kNN）分类算法的基本原理和算法流程，以及距离的计算，k值的选取等问题，然后描述了作者对k-最临近算法的具体实现。针对几组数据，分别用weka自带的k-最临近分类器和本文实现的k-最临近分类算法进行分类，分别计算出混淆矩阵并用来进行对比。最后给出了结论以及算法改进的思路和方案。
In this paper, the basic concept and procedures of k-nearest classification algorithm are introduced. The value dtermination method for k value has been discussed as well. After that the paper presents the implemention of the k-nearest algorithm. Weka's k-nearest classifier is employed alongside with this paper's implementation to analysis several dataset. The confusion matrix for each dataset is generated and compared. Finally the conclusion and improvement ideas are provided.
陈丹、窦明罡、邓泽
计算技术、计算机技术
数据挖掘k-最临近分类器机器学习
ata Miningk-NearestClassifierMachine Learning
陈丹,窦明罡,邓泽.一种k-最临近算法的设计、实现与评测[EB/OL].(2015-12-11)[2025-08-16].http://www.paper.edu.cn/releasepaper/content/201512-682.点此复制
评论