Andishe_ye Amari

Andishe_ye Amari

Fisher's linear discriminant analysis for meteorological data using the framework of Hilbert spaces with a regenerative kernel

Document Type : Original Article

Authors
1 Department of Statistics, University of Mohaghegh Ardabili
2 Department of Statistics, Shahid Beheshti University
Abstract
Nowadays, with the increasing development of science and technology, data of functional nature are easily collected. Therefore, statistical analysis of such data has become of particular importance. Like multivariate analysis, linear combinations of random variables play a key role in the analysis of functional data. Among them, the role of the theory of Hilbert spaces with a reproducing kernel is very important. In this paper, a general concept of Fisher's linear discriminant analysis, which was introduced by Shin (2008) and is a generalization of the classical multivariate method for functional data, has been reviewed. In this generalization, a two-way mapping that relates a type II random process to a Hilbert space with a reproducing kernel generated by the intraclass covariance function is used. Finally, the country's meteorological data in 2008 have been analyzed for the purpose of climate classification.
Keywords

Volume 25, Issue 2
February 2021
Pages 13-17

  • Receive Date 07 May 2025
  • First Publish Date 07 May 2025
  • Publish Date 19 February 2021