Machine Learning Approaches for Indoor Temperature Prediction in Northern Saudi Arabia   

This paper compares ten different machine learning algorithms for estimating indoor temperatures using empirical sensor readings from a real-world environment in Hail, Saudi Arabia. The data was collected during July 2023, and includes a total sample size of 8,966 records with five minutes between each reading. Five distinct environmental factors were recorded: light level, humidity, pressure, altitude and temporal features. The machine learning algorithms evaluated were Linear Regression, Ridge Regression, Lasso Regression, Elastic Net, K-Nearest Neighbors, SVR, Decision Tree, Random Forest, Gradient Boosting, and XGBoost. Each compared using R-squared, RMSE, MSE, and correlation metrics. Random Forest algorithm has achieved the highest level of performance with an R-squared value of 0.9307 and RMSE value of 0.4370 after hyperparameter tuning compared to traditional linear algorithms which have an R-squared value of below 0.33. As well, the Feature Importance analysis showed that the atmospheric pressure and humidity were the primary predictors. These findings validate tree based ensemble methods for modeling nonlinear relationships in indoor environmental systems and demonstrate practical applicability for smart building implementations in harsh climatic conditions.

[1] H. X. Zhao and F. Magoulès, “A review on the prediction of building energy consumption,” Renewable and Sustainable Energy Reviews, vol. 16, no. 6, pp. 3586–3592, 2012.
https://doi.org/10.1016/j.rser.2012.02.049

[2] A. Gassar and S. Cha, “Assessment of current energy consumption in residential buildings in Jeddah, Saudi Arabia,” Buildings, vol. 9, no. 7, p. 163, 2019.
https://doi.org/10.3390/buildings9070163

[3] A. Afram and F. Janabi-Sharifi, “Theory and applications of HVAC control systems: A review of model predictive control (MPC),” Building and Environment, vol. 72, pp. 343–355, 2014.
https://doi.org/10.1016/j.buildenv.2013.11.016

[4] K. Amasyali and N. M. El-Gohary, “A review of data-driven building energy consumption prediction studies,” Renewable and Sustainable Energy Reviews, vol. 81, pp. 1192–1205, 2018.
https://doi.org/10.1016/j.rser.2017.04.043

[5] C. Deb, F. Zhang, J. Yang, S. E. Lee, and K. W. Shah, “A review on time series forecasting techniques for building energy consumption,” Renewable and Sustainable Energy Reviews, vol. 74, pp. 902–924, 2017.
https://doi.org/10.1016/j.rser.2017.02.028

[6] C. Voyant et al., “Machine learning methods for solar radiation forecasting: A review,” Renewable Energy, vol. 105, pp. 569–582, 2017.
https://doi.org/10.1016/j.renene.2017.05.038

[7] T. Ahmad et al., “Machine learning in building energy prediction: A review,” Renewable and Sustainable Energy Reviews, vol. 144, p. 110611, 2021.
https://doi.org/10.1016/j.rser.2020.110611

[8] Y. Zhang, Z. O'Neill, B. Dong, and G. Augenbroe, “Comparisons of inverse modeling approaches for predicting building energy performance,” Building and Environment, vol. 86, pp. 177–190, 2015.
https://doi.org/10.1016/j.buildenv.2015.01.021

[9] L. Breiman, “Random forests,” Machine Learning, vol. 45, no. 1, pp. 5–32, 2001.
https://link.springer.com/article/10.1023/A:1010933404324

[10] J. H. Friedman, “Greedy function approximation: A gradient boosting machine,” Annals of Statistics, vol. 29, no. 5, pp. 1189–1232, 2001.
https://projecteuclid.org/euclid.aos/1013203451

[11] R. Zhuang et al., “Deep learning-based room temperature prediction for HVAC systems,” Sensors, vol. 23, no. 5, p. 2732, 2023.
https://doi.org/10.3390/s23052732

[12] H. Golshanfard et al., “Machine learning-based indoor temperature prediction in buildings,” Journal of Building Engineering, vol. 52, p. 104432, 2022.
https://doi.org/10.1016/j.jobe.2022.104432

[13] University of Hail Faculty, “Temporal Variations Dataset for Indoor Environmental Parameters in Northern Saudi Arabia,” 2023.
https://www.researchgate.net/publication/371764282

[14] F. Pedregosa et al., “Scikit-learn: Machine learning in Python,” Journal of Machine Learning Research, vol. 12, pp. 2825–2830, 2011.
https://jmlr.org/papers/v12/pedregosa11a.html

[15] T. Chen and C. Guestrin, “XGBoost: A scalable tree boosting system,” in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016, pp. 785–794.
https://doi.org/10.1145/2939672.2939785

[16] F. Elmaz et al., “CNN-LSTM architecture for predictive indoor temperature modeling,” Building and Environment, vol. 206, p. 108327, 2021.
https://doi.org/10.1016/j.buildenv.2021.108327

[17] N. Norouzi et al., “Applicability of deep learning algorithms for predicting indoor temperatures: Towards the development of digital twin HVAC systems,” Buildings, vol. 13, no. 6, p. 1542, 2023.
https://doi.org/10.3390/buildings13061542

[18] J. Wei et al., “Deep learning for building energy prediction: A review,” Applied Energy, vol. 272, p. 114683, 2020.
https://doi.org/10.1016/j.apenergy.2019.114683

[19] C. Fan et al., “Short-term building load prediction using machine learning methods,” Energy and Buildings, vol. 236, p. 110373, 2021.
https://doi.org/10.1016/j.enbuild.2020.110373

[20] Y. Chen et al., “Machine learning for building energy prediction: Methods and applications,” Energy, vol. 196, p. 117560, 2020.
https://doi.org/10.1016/j.energy.2020.117560

[21] G. Li et al., “Performance evaluation of sequence-to-sequence-attention model for short-term building energy prediction,” Energy, vol. 259, p. 124915, 2022.
https://doi.org/10.1016/j.energy.2022.124915