FUNDAMENTALNAYA
I PRIKLADNAYA MATEMATIKA
(FUNDAMENTAL AND APPLIED MATHEMATICS)
2000, VOLUME 6, NUMBER 2, PAGES 533-548
D. A. Pavlov
A. P. Serych
Abstract
View as HTML
View as gif image
View as LaTeX source
Excessive amount of calculation restricts possibility of using
$k$ -nearest neighbor classification algorithms. In this paper
a new method of estimation of $k$ -nearest neighbor type is
proposed. It is based on use of blocks of observed data. It
is shown that the new estimator converges in probability. Also,
the method based on the new estimator provides the same
probability of misclassification as the standard algorithm does.
But the new method requires much less calculation.
All articles are published in Russian.
Main page | Contents of the journal | News | Search |
Location: http://mech.math.msu.su/~fpm/eng/k00/k002/k00212t.htm
Last modified: September 1, 2000