Numerical modeling and forecasting temperature distribution by neural network and regression analysis

Authors

  • ADEEL TAHIR Federal Urdu University of Arts, Science & Technology, Karachi
  • MUHAMMAD ASHRAF Department of Physics, Federal Urdu University of Arts Science and Technology, Karachi
  • ZAHEER UDDIN Department of Physics, University of Karachi, Karachi
  • MUHAMMAD SARIM Department of Computer Science, Federal Urdu University of Arts Science and Technology, Karachi
  • SYED NASEEM SHAH Department of Physics, Federal Urdu University of Arts Science and Technology, Karachi

DOI:

https://doi.org/10.54302/mausam.v74i4.5513

Keywords:

Artificial Neural Network (ANN), Multi regression analysis, Daily mean temperature, Nawabshah city

Abstract

Environmental changes occur due to various parameters, and global warming is one of those parameters. It is observed that the daily mean temperature has constantly been increasing as time passes. The knowledge of temperature distribution allows us to decide the stuff that strongly depends upon temperature variation. An attempt has been made to model and forecast temperature distributions for 2018-2020. Artificial Neural Network (ANN) and multiple regression analyses have been used to forecast daily mean temperatures for one of Pakistan's cities of Sindh (Nawabshah). Environmental data from 2010 to 2020 has been used to predict daily mean temperature. The statistical errors such as RMSE, MABE and MAPE and coefficient of determination R2 are calculated to check the results' validity. Both models are suitable for predicting temperature distribution; however, ANN gives the best result. Two different regression models (linear & non-linear) are employed for the numerical fitting of temperature data; the non-linear model shows the better fitting.

Downloads

Published

01-10-2023

How to Cite

[1]
A. TAHIR, M. ASHRAF, Z. UDDIN, M. SARIM, and S. N. . SHAH, “Numerical modeling and forecasting temperature distribution by neural network and regression analysis ”, MAUSAM, vol. 74, no. 4, pp. 1183–1190, Oct. 2023.

Issue

Section

Shorter Contribution