Application of machine learning to predict daylight glare probability

Document Type : Research Paper


1 Department of Architecture, Sari Branch, Islamic Azad University, Sari, Iran

2 Department of Architecture , Sari Branch, Islamic Azad University, Sari, Iran


Daylight Glare Probability (DGP), founded on the latest glare metric, is the main challenge related to daylight glare inside buildings. Studies showed that the DGP depends on several factors, such as vertical illuminance values at the human eye factor, which is an essential parameter. In this study, we implement machine learning techniques to estimate and predict the DGP classifications, which are imperceptible, perceptible, disturbing, and intolerable based on the various influenced factors. A series of machine learning simulations have been conducted to investigate how those factors can be influenced by the degree of glare and classifications. In this research, different machine learning algorithms such as Artificial Neural Networks (multi-layer perceptron), K-Nearest Neighbors (KNN), Support Vector Machines (SVM), and Random Forest (RF) were employed to determine or predict the DGP classifications accurately. Results showed that the RF is the most effective method to classify the DGP and can predict with up to 99 % accuracy. Furthermore, the results displayed that vertical illuminance at eye level (lux), Ev, compared with other factors, has the largest influence on the DGP classifications. Consequently, machine learning is a powerful, promising, and viable option to implement in building constructions to optimize energy consumption, a global issue in the current century.


[1] M.S. Barkhordari, M.S. Es-Haghi, Straightforward prediction for responses of the concrete shear wall buildings subject to ground motions using machine learning algorithms, Int. J. Eng. 34 (2021), no. 7, 1586–1601.
[2] J. Bromley and E. Sackinger, Neural-network and k-nearest-neighbor classifiers, AT&T Bell Laboratories, (1991), 11359–910819.
[3] R.D. Clear, Discomfort glare: What do we actually know?, Lighting Res. Techno. 45 (2013), 58–141.
[4] Commission Internationale de l’Eclairage CTC 3-13, Discomfort glare in interior lighting, Proc. CIE 21st Session, CIE Publication 117, Vienna: CIE, 1995.
[5] V. Derbentsev, V. Babenko, K. Khrustalev, H. Obruch and S. Khrustalova, Comparative performance of machine learning ensemble algorithms for forecasting cryptocurrency prices, Int. J. Eng. 34 (2021), no. 1, 140–148.
[6] H. Drucker, C.J. Burges, L. Kaufman, A. Smola, and V. Vapnik, Support vector regression machines, Adv. Neural Inf. Process. Syst. 9 (1997), 155–161.
[7] H.D. Einhorn, Discomfort glare: A formula to bridge differences, Lighting Res. Technol. 11 (1979), no. 2, 90–94.
[8] V.C. Handikherkar and V.M. Phalle, Gear fault detection using machine learning techniques-a simulation-driven approach, Int. J. Eng. 34 (2021), no. 1, 212–223.
[9] T. Hastie, R. Tibshirani, J.H. Friedman and J.H. Friedman, The elements of statistical learning: Data mining, inference, and prediction, Springer, New York, 2009.
[10] M.B. Hirning, The application of luminance mapping to discomfort glare: A modified glare index for green buildings, PhD diss., Queensland University of Technology, 2014.
[11] R.G. Hopkinson, Glare from daylighting in buildings, Appl. Ergon. 3 (1972), no. 4, 206–215.
[12] J.A. Jakubiec, Validation of simplified visual discomfort calculations, Build. Simul. Optim. Conf. (BSO2018), 2018, pp. 2–11.
[13] J.A. Jakubiec and C.F. Reinhart, DIVA 2.0: Integrating daylight and thermal simulations using rhinoceros 3D, DAYSIM and EnergyPlus, Proc. Build. Simul. 12th Conf. Int. Build. Perform Simul. Assoc. 20 (2011), no. 11, 2202–2209.
[14] J.A. Jakubiec and C. Reinhart, The ’adaptive zone’–A concept for assessing discomfort glare throughout daylit spaces, Lighting Res. Technol. 44 (2012), no. 2, 149–170.
[15] J.A. Jakubiec and C.F. Reinhart, A concept for predicting occupants’ long-term visual comfort within daylit spaces, LEUKOS–J. Illum. Eng. Soc. North America, 12 (2016), no. 4, 185–202.
[16] M.G. Kent, S. Altomonte, P.R. Tregenza and R. Wilson, Temporal variables and personal factors in glare sensation, Light. Res. Technol. 48 (2016), 710–689.
[17] M. Khedmati, F. Seifi and M.J. Azizi, Time series forecasting of Bitcoin price based on autoregressive integrated moving average and machine learning approaches, Int. J. Eng. 33 (2020), no. 7, 1293–1303.
[18] I. Konstantzos, A. Tzempelikos and Y.C. Chan, Experimental and simulation analysis of daylight glare probability in offices with dynamic window shades, Build. Environ. 87 (2015), 244–254.
[19] T. Kruisselbrink, R. Dangol and A. Rosemann, Photometric measurements of lighting quality: An overview, Build. Environ. 138 (2018), 42–52.
[20] S. Kumar and G. Sahoo, A random forest classifier based on genetic algorithm for cardiovascular diseases diagnosis (research note), Int. J. Eng. 30 (2017), no. 11, 1723–1729.
[21] D. Mah, M. Kim and A. Tzempelikos, Utilization of programmable cameras for web-based sensing and control of daylight in buildings, J. Phys.: Conf. Ser. 2042 (2021), no. 1, 012114.
[22] D.S. Maitra, U. Bhattacharya and S.K. Parui, CNN based common approach to handwritten character recognition of multiple scripts, 13th Int. Conf. Doc. Anal. Recog. (ICDAR), 2015, pp. 1021–1025.
[23] A. Mentens, S. Martin, F. Descamps, J. Lataire and V.A. Jacobs, Daylight glare probability prediction for an office room, Proc. CIE Midterm Meet., 2021.
[24] T.M. Mitchell, Machine Learning, McGraw-Hill, New York, 2007.
[25] M. Namakshenas, Real-time scheduling of a flexible manufacturing system using a two-phase machine learning algorithm, Int. J. Eng. 26 (2013), no. 9, 1067–1076.
[26] A. Nazzal, O. Guler and S. Onaygil, Subjective experience of discomfort glare in a daylit computerized office in Istanbul and its mathematical prediction with the DGIN method, ARI Bull. Istanbul Tech. Univer. 54 (2005), no. 03, 96–107.
[27] B. Painter, D. Fan and J. Mardaljevic, Evidence-based daylight research: Development of a new visual comfort monitoring method, 11th Int. Light. Conf., Istanbul, Turkey, 2009.
[28] C. Pierson and M. Bodart, Validation and universalization of daylight glare probability index, LumeNet 2016, Ghent, Belgium, 2016, 82.
[29] C. Reinhart and A. Fitz, Findings from a survey on the current use of daylight simulations in building design, Energy Build. 38 (2006), no. 7, 824–835.
[30] F. Rosenblatt, Principles of neurodynamics. perceptrons and the theory of brain mechanisms, Cornell Aeronautical Lab Inc Buffalo NY, 1961.
[31] J.A.Veitch, Light, lighting, and health: Issues for consideration, LEUKOS–J. Illumin. Eng. Soc. North Amer. 2 (2005), no. 2, 85–96.
[32] M. Piechowski and A. Rowe, Building design for hot and humid climates–implications on thermal comfort and energy efficiency, IBPSA 2007–Int. Build. Perform. Simul. Assoc., 2007, pp. 122–126.
[33] G. Ward, Rendering with radiance: The art and science of lighting visualization, Morgan Kaufman, 1998.
[34] J. Wienold and J. Christoffersen, Towards a new daylight glare rating, Lux Europa, Berlin, 2005, pp. 157–161.
[35] J. Wienold and J. Christoffersen, Evaluation methods and development of a new glare prediction model for daylight environments with the use of CCD cameras, Energy Build. 38 (2006), no. 7, 743–757.
[36] J. Wienold and Fraunhofer Institute for Solar Energy Systems ISE, Daylight glare in offices, Fraunhofer Verlag, Alemanha, 2010.
[37] J. Wienold, T. Iwata, M. Sarey Khanie, E. Erell, E. Kaftan, R.G. Rodriguez, J.A. Yamin Garret´on, T. Tzempelikos, I. Konstantzos, J. Christoffersen and T.E. Kuhn, Cross-validation and robustness of daylight glare metrics, Light. Res. Technol. 51 (2019), no. 7, 983–1013.
[38] J.A. Yamin Garreton, R.G. Rodriguez, A. Ruiz and A.E. Pattini, Degree of eye-opening: A new discomfort glare indicator, Build. Environ. 88 (2015), 142–150.
Volume 15, Issue 3
March 2024
Pages 229-236
  • Receive Date: 06 January 2023
  • Revise Date: 18 March 2023
  • Accept Date: 15 April 2023