What question did this study set out to answer?

This research aims to evaluate the effectiveness of ANN and NARX models in predicting localized air temperature.

January 22, 2026Open Access

Comparative Evaluation of ANN and NARX Models for Localized Air TemperaturePrediction in Al Ain, UAE

MEMohamed H. Elnabawi NHNeveen Hamza MDMarwa Dabaieh

Key Points

This research aims to evaluate the effectiveness of ANN and NARX models in predicting localized air temperature.
Used a 12-month dataset comprising meteorological data and Typical Meteorological Year records
Trained and validated both ANN and NARX models
Compared predictive accuracy and reliability of models using R² and Mean Squared Errors
ANN model achieved R² values of 99.97% for training and 99.96% for validation and testing
NARX model exhibited higher Mean Squared Errors compared to ANN
ANN effectively captured nonlinear climatic patterns, demonstrating superior predictive reliability

Abstract

Accurate weather prediction is critical for building performance assessments and climate-responsive design, particularly in hot, arid regions. This study evaluates the effectiveness of Artificial Neural Networks (ANN) and Nonlinear AutoRegressive models with eXogenous inputs (NARX) in predicting localized air temperature for Al Ain, UAE. A 12-month dataset comprising on-site meteorological data and Typical Meteorological Year (TMY) records was used to train and validate both models. The ANN model demonstrated superior predictive accuracy, achieving R² values of 99.97% (training), 99.96% (validation), and 99.96% (testing) for the TMY dataset, and 100% across all phases for the on- site dataset. In contrast, the NARX model exhibited higher Mean Squared Errors (MSE) and lower predictive reliability. The findings highlight ANN’s capability in capturing nonlinear climatic patterns and producing reliable temperature forecasts. This study underscores the potential of ANN-based weather prediction as a tool for replacing conventional datasets that fail to reflect microclimatic variations.

AIに質問

Bookmark

View Full Paper