ORIGINAL RESEARCH
Spatial Torrential Rainfall Modelling in Pattern Analysis Based on Robust PCA Approach
 
More details
Hide details
1
Universiti Pendidikan Sultan Idris, Department of Mathematics, Faculty of Science and Mathematics, Malaysia
2
Geoinformatic Unit, Geography Section, School of Humanities, Universiti Sains Malaysia, 11800 USM, Pulau Pinang, Malaysia
3
Faculty Business and Entrepreneuship, Universiti Malaysia Kelantan, Kampus Kota, Karung Berkunci 36, Pangkalan Chepa, 16100 Kota Bharu, Kelantan, Malaysia
4
Faculty of Bioresources and Food Industry, Universiti Sultan Zainal Abidin, Besut Campus, 22200 Besut, Terengganu, Malaysia
5
Department of Mathematics and Statistics, Faculty of Applied Sciences and Technology, Universiti Tun Hussein Onn Malaysia, Malaysia
CORRESPONDING AUTHOR
Shazlyn Milleana Shaharudin   

Universiti Pendidikan Sultan Idris, Department of Mathematics, Faculty of Science and , 35900, Tanjong Malim, Malaysia
Submission date: 2020-08-20
Final revision date: 2020-11-15
Acceptance date: 2020-11-19
Online publication date: 2021-04-08
Publication date: 2021-06-09
 
Pol. J. Environ. Stud. 2021;30(4):3221–3230
 
KEYWORDS
TOPICS
ABSTRACT
In this research work, the pattern of spatial cluster had been identified for torrential rainfall data within the context of Peninsular Malaysia, which experiences heavy pour annually. Hence, a robust Principal Component Analysis (PCA) technique was employed in this study in order to address problem related to non-balance cluster(s) across patterns of rainfall stemming from skewed rainfall data. To analyze the observations made, Tukey’s biweight correlation was applied. For PCA components extraction, the optimum breakdown point was determined based on the proposed method. In order to strike a balance for extraction of number of components, as well as to hinder insignificant spatial scale or low-frequency variation, the simulation data recorded a breakdown point at 70% cumulative percentage of variance. The study outcomes revealed that the robust PCA gave better enhancement than the Pearson-based PCA did for cluster average number and quality. The findings indicate that ten rainfall patterns obtained are quite definite and clearly display the dominant role extended by the complex topography and exchange monsoons of the peninsular.
eISSN:2083-5906
ISSN:1230-1485