Adaptive Voting Rules for K-Nearest Neighbors Classifiers
Abstract
A simple form of cooperation between the k-nearest neighbors (NN) approach to classification and the neural-like property of adaptation is explored. A tunable, high level k-nearest neighbors decision rule is defined that comprehends most previous generalizations of the common majority rule. A learning procedure is developed that applies to this rule and exploits those statistical features that can be induced from the training set. The overall approach is tested on a problem of handwritten character recognition. Experiments show that adaptivity in the decision rule may improve the recognition and rejection capability of standard k-NN classifiers.
Cite
Text
Rovatti et al. "Adaptive Voting Rules for K-Nearest Neighbors Classifiers." Neural Computation, 1995. doi:10.1162/NECO.1995.7.3.594Markdown
[Rovatti et al. "Adaptive Voting Rules for K-Nearest Neighbors Classifiers." Neural Computation, 1995.](https://mlanthology.org/neco/1995/rovatti1995neco-adaptive/) doi:10.1162/NECO.1995.7.3.594BibTeX
@article{rovatti1995neco-adaptive,
title = {{Adaptive Voting Rules for K-Nearest Neighbors Classifiers}},
author = {Rovatti, Riccardo and Ragazzoni, R. and Kovács-Vajna, Zsolt Miklós and Guerrieri, Roberto},
journal = {Neural Computation},
year = {1995},
pages = {594-605},
doi = {10.1162/NECO.1995.7.3.594},
volume = {7},
url = {https://mlanthology.org/neco/1995/rovatti1995neco-adaptive/}
}