An Improved KNN Classification Algorithm based on Sampling

DOI

10.2991/ammee-17.2017.45How to use a DOI?

Keywords

KNN, classification algorithm, computational overhead, sampling.

Abstract

K nearest neighbor (KNN) algorithm has been widely used as a simple and effective classification algorithm. The traditional KNN classification algorithm will find k nearest neighbors, it is necessary to calculate the distance from the test sample to all training samples. When the training sample data is very large, it will produce a high computational overhead, resulting in a decline in classification speed. Therefore, we optimize the distance calculation of the KNN algorithm. Since KNN only considers the k samples of the shortest distance from the test sample to the nearest training sample point, the large distance training has no effect on the classification of the algorithm. The improved method is to sample the training data around the test data, which reduces the number of distance calculation of the test data to each training data, and reduces the time complexity of the algorithm. The experimental results show that the optimized KNN classification algorithm is superior to the traditional KNN algorithm.

Copyright

© 2017, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Zhiwei Cheng
AU  - Caisen Chen
AU  - Xuehuan Qiu
AU  - Huan Xie
PY  - 2017/06
DA  - 2017/06
TI  - An Improved KNN Classification Algorithm based on Sampling
BT  - Proceedings of the Advances in Materials, Machinery, Electrical Engineering (AMMEE 2017)
PB  - Atlantis Press
SP  - 220
EP  - 225
SN  - 2352-5401
UR  - https://doi.org/10.2991/ammee-17.2017.45
DO  - 10.2991/ammee-17.2017.45
ID  - Cheng2017/06
ER  -