On Parzen windows classifiers

Jing Peng, Guna Seetharaman

Research output: Contribution to journalConference articleResearchpeer-review

Abstract

Parzen Windows classifiers have been applied to a variety of density estimation as well as classification tasks with considerable success. Parzen Windows are known to converge in the asymptotic limit. However, there is a lack of theoretical analysis on their performance with finite samples. In this paper we show a connection between Parzen Windows and the regularized least squares algorithm, which has a well-established foundation in computational learning theory. This connection allows us to provide useful insight into Parzen Windows classifiers and their performance in finite sample settings. Finally, we show empirical results on the performance of Parzen Windows classifiers using a number of real data sets.

Original languageEnglish
Article number7041924
JournalProceedings - Applied Imagery Pattern Recognition Workshop
Volume2015-February
Issue numberFebruary
DOIs
StatePublished - 12 Feb 2015
Event2014 IEEE Applied Imagery Pattern Recognition Workshop, AIPR 2014 - Washington, United States
Duration: 14 Oct 201416 Oct 2014

Fingerprint

Classifiers

Cite this

Peng, Jing ; Seetharaman, Guna. / On Parzen windows classifiers. In: Proceedings - Applied Imagery Pattern Recognition Workshop. 2015 ; Vol. 2015-February, No. February.
@article{b361c2c6f277454dacca2c411e36aa05,
title = "On Parzen windows classifiers",
abstract = "Parzen Windows classifiers have been applied to a variety of density estimation as well as classification tasks with considerable success. Parzen Windows are known to converge in the asymptotic limit. However, there is a lack of theoretical analysis on their performance with finite samples. In this paper we show a connection between Parzen Windows and the regularized least squares algorithm, which has a well-established foundation in computational learning theory. This connection allows us to provide useful insight into Parzen Windows classifiers and their performance in finite sample settings. Finally, we show empirical results on the performance of Parzen Windows classifiers using a number of real data sets.",
author = "Jing Peng and Guna Seetharaman",
year = "2015",
month = "2",
day = "12",
doi = "10.1109/AIPR.2014.7041924",
language = "English",
volume = "2015-February",
journal = "Proceedings - Applied Imagery Pattern Recognition Workshop",
issn = "2164-2516",
number = "February",

}

On Parzen windows classifiers. / Peng, Jing; Seetharaman, Guna.

In: Proceedings - Applied Imagery Pattern Recognition Workshop, Vol. 2015-February, No. February, 7041924, 12.02.2015.

Research output: Contribution to journalConference articleResearchpeer-review

TY - JOUR

T1 - On Parzen windows classifiers

AU - Peng, Jing

AU - Seetharaman, Guna

PY - 2015/2/12

Y1 - 2015/2/12

N2 - Parzen Windows classifiers have been applied to a variety of density estimation as well as classification tasks with considerable success. Parzen Windows are known to converge in the asymptotic limit. However, there is a lack of theoretical analysis on their performance with finite samples. In this paper we show a connection between Parzen Windows and the regularized least squares algorithm, which has a well-established foundation in computational learning theory. This connection allows us to provide useful insight into Parzen Windows classifiers and their performance in finite sample settings. Finally, we show empirical results on the performance of Parzen Windows classifiers using a number of real data sets.

AB - Parzen Windows classifiers have been applied to a variety of density estimation as well as classification tasks with considerable success. Parzen Windows are known to converge in the asymptotic limit. However, there is a lack of theoretical analysis on their performance with finite samples. In this paper we show a connection between Parzen Windows and the regularized least squares algorithm, which has a well-established foundation in computational learning theory. This connection allows us to provide useful insight into Parzen Windows classifiers and their performance in finite sample settings. Finally, we show empirical results on the performance of Parzen Windows classifiers using a number of real data sets.

UR - http://www.scopus.com/inward/record.url?scp=84937907295&partnerID=8YFLogxK

U2 - 10.1109/AIPR.2014.7041924

DO - 10.1109/AIPR.2014.7041924

M3 - Conference article

VL - 2015-February

JO - Proceedings - Applied Imagery Pattern Recognition Workshop

JF - Proceedings - Applied Imagery Pattern Recognition Workshop

SN - 2164-2516

IS - February

M1 - 7041924

ER -