Modeling Universal Thermal Climate Index Thermal Stress in Iran’s Hot Zones Using Neural Networks and Naïve Bayes

Aim: This study aimed to assess the impact of heat stressors in various environments across the expansive hot regions of Iran, with the goal of supporting workforce health and well-being. Methods: This descriptive-analytical study was conducted in four main stages: (1) identifying and measuring fact...

Full description

Saved in:
Bibliographic Details
Main Authors: Sajad Zare, Reza Esmaeili, Rasoul Hemmatjo
Format: Article
Language:English
Published: Wolters Kluwer Medknow Publications 2025-05-01
Series:International Journal of Environmental Health Engineering
Subjects:
Online Access:https://journals.lww.com/10.4103/ijehe.ijehe_21_24
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850123274204217344
author Sajad Zare
Reza Esmaeili
Rasoul Hemmatjo
author_facet Sajad Zare
Reza Esmaeili
Rasoul Hemmatjo
author_sort Sajad Zare
collection DOAJ
description Aim: This study aimed to assess the impact of heat stressors in various environments across the expansive hot regions of Iran, with the goal of supporting workforce health and well-being. Methods: This descriptive-analytical study was conducted in four main stages: (1) identifying and measuring factors influencing the Universal Thermal Climate Index (UTCI), (2) calculating the UTCI as the dependent variable, (3) applying neural network (NN) and Naïve Bayes (NB) algorithms to weigh and model the effective factors on the UTCI, and (4) evaluating the accuracy of each model. Data modeling utilized Python’s scikit-learn package (version 3.7) and the Orange toolkit. Results: The average UTCI values in hot-dry and hot-humid regions were 30.49 and 39.48, respectively. In models for hot-dry regions, both algorithms identified dry temperature (Ta) as a significant factor. For hot-humid regions, the NB algorithm identified mean radiant temperature (Tmrt) as the primary factor, while the NN algorithm highlighted dry temperature (Ta). Model accuracy ranged from 74% to 94%, with NN algorithms demonstrating higher accuracy compared to NB algorithms. Conclusion: In the models for hot-dry regions, both algorithms predicted Ta and then Tmrt as the principal factors. For hot-humid regions, it was inferred that Tmrt is the main influencing factor on the UTCI.
format Article
id doaj-art-4d1a2f6bec3b436fbc71af96dbf06662
institution OA Journals
issn 2277-9183
language English
publishDate 2025-05-01
publisher Wolters Kluwer Medknow Publications
record_format Article
series International Journal of Environmental Health Engineering
spelling doaj-art-4d1a2f6bec3b436fbc71af96dbf066622025-08-20T02:34:39ZengWolters Kluwer Medknow PublicationsInternational Journal of Environmental Health Engineering2277-91832025-05-011429910.4103/ijehe.ijehe_21_24Modeling Universal Thermal Climate Index Thermal Stress in Iran’s Hot Zones Using Neural Networks and Naïve BayesSajad ZareReza EsmaeiliRasoul HemmatjoAim: This study aimed to assess the impact of heat stressors in various environments across the expansive hot regions of Iran, with the goal of supporting workforce health and well-being. Methods: This descriptive-analytical study was conducted in four main stages: (1) identifying and measuring factors influencing the Universal Thermal Climate Index (UTCI), (2) calculating the UTCI as the dependent variable, (3) applying neural network (NN) and Naïve Bayes (NB) algorithms to weigh and model the effective factors on the UTCI, and (4) evaluating the accuracy of each model. Data modeling utilized Python’s scikit-learn package (version 3.7) and the Orange toolkit. Results: The average UTCI values in hot-dry and hot-humid regions were 30.49 and 39.48, respectively. In models for hot-dry regions, both algorithms identified dry temperature (Ta) as a significant factor. For hot-humid regions, the NB algorithm identified mean radiant temperature (Tmrt) as the primary factor, while the NN algorithm highlighted dry temperature (Ta). Model accuracy ranged from 74% to 94%, with NN algorithms demonstrating higher accuracy compared to NB algorithms. Conclusion: In the models for hot-dry regions, both algorithms predicted Ta and then Tmrt as the principal factors. For hot-humid regions, it was inferred that Tmrt is the main influencing factor on the UTCI.https://journals.lww.com/10.4103/ijehe.ijehe_21_24hot dryhot humiduniversal thermal climate indexweighting
spellingShingle Sajad Zare
Reza Esmaeili
Rasoul Hemmatjo
Modeling Universal Thermal Climate Index Thermal Stress in Iran’s Hot Zones Using Neural Networks and Naïve Bayes
International Journal of Environmental Health Engineering
hot dry
hot humid
universal thermal climate index
weighting
title Modeling Universal Thermal Climate Index Thermal Stress in Iran’s Hot Zones Using Neural Networks and Naïve Bayes
title_full Modeling Universal Thermal Climate Index Thermal Stress in Iran’s Hot Zones Using Neural Networks and Naïve Bayes
title_fullStr Modeling Universal Thermal Climate Index Thermal Stress in Iran’s Hot Zones Using Neural Networks and Naïve Bayes
title_full_unstemmed Modeling Universal Thermal Climate Index Thermal Stress in Iran’s Hot Zones Using Neural Networks and Naïve Bayes
title_short Modeling Universal Thermal Climate Index Thermal Stress in Iran’s Hot Zones Using Neural Networks and Naïve Bayes
title_sort modeling universal thermal climate index thermal stress in iran s hot zones using neural networks and naive bayes
topic hot dry
hot humid
universal thermal climate index
weighting
url https://journals.lww.com/10.4103/ijehe.ijehe_21_24
work_keys_str_mv AT sajadzare modelinguniversalthermalclimateindexthermalstressiniranshotzonesusingneuralnetworksandnaivebayes
AT rezaesmaeili modelinguniversalthermalclimateindexthermalstressiniranshotzonesusingneuralnetworksandnaivebayes
AT rasoulhemmatjo modelinguniversalthermalclimateindexthermalstressiniranshotzonesusingneuralnetworksandnaivebayes