Title: Modelling and analysing the daily temperature of several cities using mixture Gaussian distributions

Authors: Zuhair A. Al-Hemyari; Jamal N. Al Abbasi

Addresses: Department of Mathematical and Physical Sciences, CAS, University of Nizwa, P.O. Box 33, PC 616, Nizwa, Oman ' Department of Statistics, AL-Nahrain University, Baghdad, Iraq

Abstract: The rise in the surface temperature of the Earth since the 19th century is a clear indication of the negative and significant impact humans have had upon it. One of the results of this is that many researchers have studied meteorological problems. Some of the important applications for modelling the daily maximum and minimum temperatures reveal the importance of modelling the data for these phenomena using mixed distributions. In order to model the data of three cities over 40 years of daily maximum/minimum temperatures, heavy statistical computations have been carried out, and several single and mixture Gaussian distributions have developed the modelling of data of three cities for the period of 1980-2019 of daily maximum/minimum temperatures. In addition, the computations of expectation maximisation algorithm estimators, the tests of goodness, and the frequency distributions have been developed to carry out accurate analyses. The data is analysed, several indicators are seized, and the conclusions, recommendations, limitations, and future research directions are discussed.

Keywords: Gaussian distribution; mixture distributions; MLE; EM algorithms; goodness of fit; Kolmogorov-Smirnov test; Anderson-Darling test; daily temperatures.

DOI: 10.1504/IJCSM.2023.131630

International Journal of Computing Science and Mathematics, 2023 Vol.17 No.4, pp.320 - 341

Received: 24 Jul 2021
Received in revised form: 31 Mar 2022
Accepted: 13 Sep 2022
Published online: 21 Jun 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article