Instances selection for NN with fuzzy rough technique
Instances selection for NN with fuzzy rough technique
复制标题
DOI：
10.1109/icmlc.2011.6016939
复制
发表时间：
2011-07
期刊：
影响因子：
--
通讯作者：
Xiao-Meng Kang;Xiao-Peng Liu;Jun-Hai Zhai;Meng-Yao Zhai
中科院分区：
文献类型：
--
作者：
Xiao-Meng Kang;Xiao-Peng Liu;Jun-Hai Zhai;Meng-Yao Zhai
NN 算法是一种简单且众所周知的监督学习方案，它通过在训练集中找到其最近邻来对未见过的实例进行分类。神经网络的主要缺点是整个训练集必须存储在计算机中才能对未见过的实例进行分类。为了解决这个问题，P. Hart 提出了压缩最近邻（CNN）算法。然而，CNN从整个训练集中选择重要的实例，与NN一样存在需要大量内存的问题。在本文中，我们提出了一种使用模糊粗糙技术从边界区域选择实例的算法。实验结果证明了我们提出的方法的有效性。
The NN algorithm is a simple and well-known supervised learning scheme which classifies an unseen instance by finding its closest neighbor in training set. The main drawback of NN is that the whole training set must be stored in the computer to classify an unseen instance. In order to deal with this problem, P. Hart proposed the condensed nearest neighbor (CNN) algorithm. However, CNN select the important instances from the whole training set, which suffers from the problem of large memory requirement same as NN. In this paper, we propose an algorithm to select instances from the border region with fuzzy rough technique. The experimental results demonstrate the effectiveness of our proposed method.