Tilted Nadaraya-Watson regression estimator

Mohammad T. Shakeri, Farzaneh Boroumand*, Hassan Doosti, Nino Kordzakhia, Mahdi Salehi

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference proceeding contributionpeer-review

1 Citation (Scopus)

Abstract

In nonparametric statistics the tilting techniques are sustainably used for adjusting an empirical distribution by replacing uniform distribution of weights by general multinomial distribution. In this paper a tilting approach has been used for minimizing “the distance” to an infinite order (IO) regression estimator, a comparator regression function estimator. We also provide the simulation study results illustrating the tilted version of the Nadaraya-Watson (N-W) estimator performs better than its classical analog (the N-W estimator) in terms of Median Integrated Squared Error (MISE). In addition, the performance of the tilted N-W regression function estimator has been examined using the Italy’s COVID-19 deaths data.

Original languageEnglish
Title of host publication13th Chaotic Modeling and Simulation International Conference
EditorsChristos H. Skiadas, Yiannis Dimotikalis
Place of PublicationCham
PublisherSpringer, Springer Nature
Pages797-803
Number of pages7
ISBN (Electronic)9783030707958
ISBN (Print)9783030707941, 9783030707972
DOIs
Publication statusPublished - 2021
Event13th Chaotic Modeling and Simulation International Conference, CHAOS 2020 - Florence, Italy
Duration: 9 Jun 202012 Jun 2020

Publication series

NameSpringer Proceedings in Complexity
PublisherSpringer
ISSN (Print)2213-8684
ISSN (Electronic)2213-8692

Conference

Conference13th Chaotic Modeling and Simulation International Conference, CHAOS 2020
Country/TerritoryItaly
CityFlorence
Period9/06/2012/06/20

Keywords

  • Nadaraya-Watson estimator
  • Tilted Nadaraya-Watson estimator
  • Infinite order estimator
  • Kernel estimator
  • Trapezoidal kernel
  • Cross-validation function
  • MISE
  • ISE

Fingerprint

Dive into the research topics of 'Tilted Nadaraya-Watson regression estimator'. Together they form a unique fingerprint.

Cite this