A Training Set Size Reduction Approach Based on Covering Scheme
The nearest neighbor (NN) rule is one of the most popular nonparametric pattern classification algorithms. For large data sets, the computational cost of classifying pattern process using NN can be expensive. A way to alleviate this problem is through the reducing training set size approach. This means we remove patterns that are more of a computational burden but do not contribute to better classification accuracy. In this paper, a new training set size reduction method is presented. The proposed idea is based on defining the so-called covering sets and the corresponding representatives. In the classification stage, we need only take into account the representatives which are a part of the training samples. Experimental results show that the proposed approach effectively reduces the number of prototypes while maintaining the same level of classification accuracy as the traditional NN.
Template reduction Covering rule Nearest neighbor (NN)
Jinfu Yang Min Song Mingai Li
School of Electronic Information and Control Engineering, Beijing University of Technology, Beijing, 100124, China
国际会议
2010 International Conference on Circuit and Signal Processing(2010年电路与信号处理国际会议 ICCSP 2010)
上海
英文
361-364
2010-12-25（万方平台首次上网日期，不代表论文的发表时间）