Functional Data Analysis for Extreme Data

Noor Izyan Mohamad Adnan; Mohd  Bakri Adam; Mohd  Yusoff Ishak; Noor Akma Ibrahim  and Mohammad Noor Amal Azmai

doi:10.17485/ijst/2016/v9i28/97356

Article

Functional Data Analysis for Extreme Data

VIEWS 896
PDF 315

Abstract
Full-Text HTML
Full-Text PDF
How to Cite

Indian Journal of Science and Technology

DOI: 10.17485/ijst/2016/v9i28/97356

Year: 2016, Volume: 9, Issue: 28, Pages: 1-6

Original Article

Functional Data Analysis for Extreme Data

Noor Izyan Mohamad Adnan^{1 *}, Mohd. Bakri Adam¹ , Mohd. Yusoff Ishak² , Noor Akma Ibrahim³ and Mohammad Noor Amal Azmai⁴

¹Institute for Mathematical Research, [email protected]
²Faculty of Environmental Studies,
³Mathematical Department, Faculty of Science,
⁴ Depatment of Biology, Faculty of Science,
*Author for correspondence
Noor Izyan Mohamad Adnan
Institute for Mathematical Research,
Email: [email protected]

This work is licensed under a Creative Commons Attribution 4.0 International License.

Abstract

The performance of extreme data is observed by using functional data analysis with two extreme values theory approaches. Functional data analysis is one of the techniques to represent data in a functional form or as a smooth curve rather than in a discrete form. This functional observation will be fitted using fourier series by least squares and roughness penalty method. The data will be tested on block maxima and r-largest order statistics approaches to indicate what numbers of data required to have the best fitted curve. The finding illustrates three r-largest order statistics approach gives a better performance for functional data analysis which deals with extreme values data.
Keywords: Functional Data Analysis, Extreme Values Theory, Fourier Series, Generalize Cross-validation