A Simple Noise-Tolerant Abstraction Algorithm for Fast k-NN Classification
A Simple Noise-Tolerant Abstraction Algorithm for Fast k-NN Classification
复制标题
DOI：
10.1007/978-3-642-28931-6_20
复制
发表时间：
2012-03
期刊：
影响因子：
--
通讯作者：
Stefanos Ougiaroglou;Georgios Evangelidis
中科院分区：
文献类型：
--
作者：
Stefanos Ougiaroglou;Georgios Evangelidis
k-近邻分类器是一种应用广泛且有效的分类方法。maink-NN的缺点是，当应用于大型数据集时，它涉及高计算成本。为了加快分类过程，已经提出了许多数据简化技术。然而，它们的有效性取决于数据中的噪声水平。本文证明了k-均值聚类算法可以作为一种抗噪声的数据约简技术。所进行的实验研究表明，如果减少的数据集包括k均值质心作为初始数据的代表，则性能不会受到噪声的负面影响。
Thek-Nearest Neighbor (k-NN) classifier is a widely-used and effective classification method. The maink-NN drawback is that it involves high computational cost when applied on large datasets. Many Data Reduction Techniques have been proposed in order to speed-up the classification process. However, their effectiveness depends on the level of noise in the data. This paper shows that thek-means clustering algorithm can be used as a noise-tolerant Data Reduction Technique. The conducted experimental study illustrates that if the reduced dataset includes the k-means centroids as representatives of the initial data, performance is not negatively affected as much by the addition of noise.