1099-128X/asset/CEM_left.gif?v=1&s=bf7a32b94d86cfd950babd255fbe81e66d033e4b)
1099-128X/asset/CEM_right.gif?v=1&s=4630211ecefb8b6241dad7b782e7b742d7a9891a)
Short Communication
Variable scaling and nearest neighbor methods
Article first published online: 23 DEC 2008
DOI: 10.1002/cem.1211
Copyright © 2008 John Wiley & Sons, Ltd.
1099-128X/asset/cover.gif?v=1&s=2e3045c3733baa4258989f44bd61b29dd74ee736 "Journal of Chemometrics")
Additional Information
How to Cite
Gertheiss, J. and Tutz, G. (2009), Variable scaling and nearest neighbor methods. Journal of Chemometrics, 23: 149–151. doi: 10.1002/cem.1211
Publication History
- Issue published online: 3 MAR 2009
- Article first published online: 23 DEC 2008
- Manuscript Accepted: 18 NOV 2008
- Manuscript Revised: 17 NOV 2008
- Manuscript Received: 8 AUG 2008
- Abstract
- References
- Cited By
Keywords:
- classification;
- k-nearest-neighbors;
- standardizing predictors
When employing nearest neighbor classifiers scaling of input variables is often useful. In this paper we propose a small modification in usual data preprocessing: scaling of variables should be done by use of pooled variances instead of overall ones. Thus prediction accuracy is distinctly improved in some situations.
Abstract
When employing nearest neighbor classifiers scaling of input variables is often useful. In this paper we propose a small modification in usual data preprocessing: scaling of variables should be done by use of pooled variances instead of overall ones. Thus prediction accuracy is distinctly improved in some situations. Copyright © 2008 John Wiley & Sons, Ltd.