Analytical and Applied Study for Estimating the Heat Index in Gharyan City Using Artificial Intelligence Techniques

Abstract

This research aims to study and analyze the Heat Index in Gharyan city, located in the Western Mountain region of Libya, used on climate data spanning a long period from 1995 to 2024. The index was calculated using the U.S. National Weather Service (NWS) equation for monthly temperature and humidity. The results showed that the summer months (June–September) are characterized a perceived temperature that exceeds the actual temperature, with the maximum difference reaching +0.8°C in August, highlighting the importance of this index in assessing thermal stress on the population. Five artificial intelligence models were developed and trained to predict the Heat Index: Linear Regression, Ridge Regression, Lasso Regression, Random Forest, and Neural Networks. The results demonstrated the superiority of the Lasso Regression model, achieving high prediction accuracy (R² = 0.9990, RMSE = 0.1554°C), making it an effective tool for rapid prediction. Additionally, the research provides practical recommendations to support urban planning, organize daily activities, and develop early warning systems for heat waves. It also lays the groundwork for roader future studies on thermal comfort in Libya amidst accelerating climate change.