Building energy consumption has been catching the world’s attention, due to its significant contribution to the global total energy consumption. Many ways explored to conserve building energy consumption. One crucial early stage for choosing the practical policies to conserve energy consumption is capturing the energy consumption data first. Capturing energy consumption data may lead to broader insight. For example, what device consumes most of the electricity and re-evaluate whether energy consumption has possibilities to be conserved. Thus, with the data, benchmarking and simulation are the example with a sophisticated method and framework, HVAC optimization, retrofit recommendation, equipment replacement, and maintenance recommendation.
In some cases, like a hotel building or university building, room re-arrangement could be one of the solutions to conserve the energy. Hence, the energy consumption data analysis is critical for determining the data-driven energy conservation method. The data-driven method has been prevalent in data analytics research because it helps to remove assumptions that could create bias while the model applied to real-world situations.
In this research, the primary purpose is to presents an applicable framework for energy conservation in the early stages. The framework applied in the university building complex data. BGCP and LRTC method, along with tensor-shaped data, applied to impute the missing value data. BGCP and LRTC performance are quite well when it comes to imputing the random-missing data. However, to deal with the whole day missing data, LSTM sequence-to-sequence is presented to fill in the missing value. To deal with multi-prediction, which is the area that is currently explored to help multi-building stakeholder, deep-learning algorithms utilized. This multi-building prediction is a multivariate time-series supervised problem to predict energy consumption for each building 24-hours ahead. The settings of the input, output shape, and layer arrangement of the deep learning examined to improve the accuracy of the model. The comparison of deep learning model CNN, LSTM, the combination of the two methods also integrate clustering result to the model will be presented in this research.
Finally, the proposed framework for further implementation presented to give an illustration of possible energy conservation framework in the university building complex. The further data suggestions and advantages for the framework improvement also presented to give insight about the use of the data.

Building energy consumption has been catching the world’s attention, due to its significant contribution to the global total energy consumption. Many ways explored to conserve building energy consumption. One crucial early stage for choosing the practical policies to conserve energy consumption is capturing the energy consumption data first. Capturing energy consumption data may lead to broader insight. For example, what device consumes most of the electricity and re-evaluate whether energy consumption has possibilities to be conserved. Thus, with the data, benchmarking and simulation are the example with a sophisticated method and framework, HVAC optimization, retrofit recommendation, equipment replacement, and maintenance recommendation.
In some cases, like a hotel building or university building, room re-arrangement could be one of the solutions to conserve the energy. Hence, the energy consumption data analysis is critical for determining the data-driven energy conservation method. The data-driven method has been prevalent in data analytics research because it helps to remove assumptions that could create bias while the model applied to real-world situations.
In this research, the primary purpose is to presents an applicable framework for energy conservation in the early stages. The framework applied in the university building complex data. BGCP and LRTC method, along with tensor-shaped data, applied to impute the missing value data. BGCP and LRTC performance are quite well when it comes to imputing the random-missing data. However, to deal with the whole day missing data, LSTM sequence-to-sequence is presented to fill in the missing value. To deal with multi-prediction, which is the area that is currently explored to help multi-building stakeholder, deep-learning algorithms utilized. This multi-building prediction is a multivariate time-series supervised problem to predict energy consumption for each building 24-hours ahead. The settings of the input, output shape, and layer arrangement of the deep learning examined to improve the accuracy of the model. The comparison of deep learning model CNN, LSTM, the combination of the two methods also integrate clustering result to the model will be presented in this research.
Finally, the proposed framework for further implementation presented to give an illustration of possible energy conservation framework in the university building complex. The further data suggestions and advantages for the framework improvement also presented to give insight about the use of the data.

Abdi, H., & Williams, L. J. (2010). Tukey’s honestly significant difference (HSD) test. Encyclopedia of research design, 3, 583-585.
Altan, H., Douglas, J., & Kim, Y. (2014). Energy performance analysis of university buildings: Case studies at Sheffield University, UK. J Architectural Engineering Technology, 3(129), 2.
Andersen, F. M., Larsen, H. V., & Boomsma, T. K. (2013). Long-term forecasting of hourly electricity load: Identification of consumption profiles and segmentation of customers. Energy conversion and Management, 68, 244-252.
Berglund, P., & Heeringa, S. G. (2014). Multiple imputation of missing data using SAS: SAS Institute.
Bureau, C. W. (2020). Retrieved from http://e-service.cwb.gov.tw/
Chang, Y., Kim, C. S., Miller, J. I., Park, J. Y., & Park, S. (2016). A new approach to modeling the effects of temperature fluctuations on monthly electricity demand. Energy Economics, 60, 206-216.
Chen, F. (2009). Bayesian modeling using the MCMC procedure. Paper presented at the Proceedings of the SAS Global Forum 2008 Conference, Cary NC: SAS Institute Inc.
Chen, S., Zhang, X., Wei, S., Yang, T., Guan, J., Yang, W., . . . Xu, Y. (2019). An energy planning oriented method for analyzing spatial-temporal characteristics of electric loads for heating/cooling in district buildings with a case study of one university campus. Sustainable Cities and Society, 51, 101629.
Chen, X., He, Z., & Sun, L. (2019). A Bayesian tensor decomposition approach for spatiotemporal traffic data imputation. Transportation research part C: emerging technologies, 98, 73-84.
Chen, X., Yang, J., & Sun, L. (2020). A nonconvex low-rank tensor completion model for spatiotemporal traffic data imputation. Transportation research part C: emerging technologies, 117, 102673.
Chen, Y., Tong, Z., Zheng, Y., Samuelson, H., & Norford, L. (2020). Transfer learning with deep neural networks for model predictive control of HVAC and natural ventilation in smart buildings. Journal of cleaner production, 254, 119866.
Chung, M. H., & Rhee, E. K. (2014). Potential opportunities for energy conservation in existing buildings on university campus: A field survey in Korea. Energy and Buildings, 78, 176-182.
Deb, C., Zhang, F., Yang, J., Lee, S. E., & Shah, K. W. (2017). A review on time series forecasting techniques for building energy consumption. Renewable and Sustainable Energy Reviews, 74, 902-924.
Dong, X., Qian, L., & Huang, L. (2017). Short-term load forecasting in smart grid: A combined CNN and K-means clustering approach. Paper presented at the 2017 IEEE International Conference on Big Data and Smart Computing (BigComp).
Du Prel, J.-B., Hommel, G., Röhrig, B., & Blettner, M. (2009). Confidence interval or p-value?: part 4 of a series on evaluation of scientific publications. Deutsches Ärzteblatt International, 106(19), 335.
Enders, W. (2008). Applied econometric time series: John Wiley & Sons.
Fathi, S., & Srinivasan, R. S. (2015). Analysis of energy performance of university campus buildings using statistical and energy modeling approaches. Paper presented at the 2015 Winter Simulation Conference (WSC).
Fawaz, H. I., Forestier, G., Weber, J., Idoumghar, L., & Muller, P.-A. (2019). Deep learning for time series classification: a review. Data Mining and Knowledge Discovery, 33(4), 917-963.
Flores, J. H. F., Engel, P. M., & Pinto, R. C. (2012). Autocorrelation and partial autocorrelation functions to improve neural networks models on univariate time series forecasting. Paper presented at the The 2012 International Joint Conference on Neural Networks (IJCNN).
Ghadban, M., Baayoun, A., Lakkis, I., Najem, S., Saliba, N. A., & Shihadeh, A. (2020). A novel method to improve temperature forecast in data-scarce urban environments with application to the Urban Heat Island in Beirut. Urban Climate, 33, 100648.
Global Alliance for Buildings and Construction, I. E. A., and United Nations Environment Programme. (2019). 2019 Global Status Report for Buildings and Construction : Towards a zero-emission, efficient and resilient buildings and construction sector ( 978-92-807-3768-4 ). Retrieved from
Greff, K., Srivastava, R. K., Koutník, J., Steunebrink, B. R., & Schmidhuber, J. (2016). LSTM: A search space odyssey. IEEE transactions on neural networks and learning systems, 28(10), 2222-2232.
Guan, L. (2009). Implication of global warming on air-conditioned office buildings in Australia. Building Research & Information, 37(1), 43-54.
Guerrieri, M., La Gennusa, M., Peri, G., Rizzo, G., & Scaccianoce, G. (2019). University campuses as small-scale models of cities: Quantitative assessment of a low carbon transition path. Renewable and Sustainable Energy Reviews, 113, 109263.
Harish, S., Singh, N., & Tongia, R. (2020). Impact of temperature on electricity demand: Evidence from Delhi and Indian states. Energy Policy, 140, 111445.
Hawkins, D., Hong, S., Raslan, R., Mumovic, D., & Hanna, S. (2012). Determinants of energy use in UK higher education buildings using statistical and artificial neural network methods. International Journal of Sustainable Built Environment, 1(1), 50-63.
Huovila, A., Tuominen, P., & Airaksinen, M. (2017). Effects of building occupancy on indicators of energy efficiency. Energies, 10(5), 628.
IEA. (2019). Energy Efficiency 2019. Retrieved from Paris: https://www.iea.org/reports/energy-efficiency-2019
IEA. (2020a). Cooling. Retrieved from https://www.iea.org/fuels-and-technologies/cooling
IEA. (2020b). Tracking Buildings 2020. Retrieved from Paris: https://www.iea.org/reports/tracking-buildings-2020
Jana, R. K., Ghosh, I., & Sanyal, M. K. (2020). A granular deep learning approach for predicting energy consumption. Applied Soft Computing, 89, 106091.
Khan, S., Javaid, N., Chand, A., Khan, A. B. M., Rashid, F., & Afridi, I. U. (2019). Electricity load forecasting for each day of week using deep CNN. Paper presented at the Workshops of the International Conference on Advanced Information Networking and Applications.
Kim, T., Kang, B., Kim, H., Park, C., & Hong, W.-H. (2019). The study on the Energy Consumption of middle school facilities in Daegu, Korea. Energy Reports, 5, 993-1000.
Kneifel, J., & Webb, D. (2016). Predicting energy performance of a net-zero energy building: A statistical approach. Applied Energy, 178, 468-483.
Kolokotroni, M., Ren, X., Davies, M., & Mavrogianni, A. (2012). London's urban heat island: Impact on current and future energy consumption in office buildings. Energy and Buildings, 47, 302-311.
Kontokosta, C. E., & Tull, C. (2017). A data-driven predictive model of city-scale energy use in buildings. Applied Energy, 197, 303-317.
Lin, G., & Claridge, D. E. (2015). A temperature-based approach to detect abnormal building energy consumption. Energy and Buildings, 93, 110-118.
Lourenço, P., Pinheiro, M. D., & Heitor, T. (2019). Light use patterns in Portuguese school buildings: User comfort perception, behaviour and impacts on energy consumption. Journal of cleaner production, 228, 990-1010.
Luo, R., Han, Y., & Zhou, X. (2017). Characteristics of campus energy consumption in North China University of Science and Technology. Procedia Engineering, 205, 3816-3823.
Madathil, B., Sagheer, S. V. M., Rahiman V, A., Tom, A. J., Francis, J., & George, S. N. (2019). Tensor Low Rank Modeling and Its Applications in Signal Processing. arXiv preprint arXiv:1912.03435.
Marinakis, V., Doukas, H., Tsapelas, J., Mouzakitis, S., Sicilia, Á., Madrazo, L., & Sgouridis, S. (2018). From big data to smart energy services: An application for intelligent energy management. Future Generation Computer Systems.
Montgomery, D. C. (2007). Introduction to statistical quality control: John Wiley & Sons.
NTUST. (2020a). Academic Calendar. Retrieved from https://www.academic.ntust.edu.tw/
NTUST. (2020b). Course Query System. Retrieved from https://querycourse.ntust.edu.tw/querycourse/
Perera, A., Nik, V. M., Chen, D., Scartezzini, J.-L., & Hong, T. (2020). Quantifying the impacts of future climate variations and extreme climate events on energy systems.
Rai, P., Wang, Y., Guo, S., Chen, G., Dunson, D., & Carin, L. (2014). Scalable Bayesian low-rank decomposition of incomplete multiway tensors. Paper presented at the International Conference on Machine Learning.
Rizwan, A. M., Dennis, L. Y., & Chunho, L. (2008). A review on the generation, determination and mitigation of Urban Heat Island. Journal of Environmental Sciences, 20(1), 120-128.
Rutherford, A. (2001). Introducing ANOVA and ANCOVA: a GLM approach: Sage.
Salakhutdinov, R., & Mnih, A. (2008). Bayesian probabilistic matrix factorization using Markov chain Monte Carlo. Paper presented at the Proceedings of the 25th international conference on Machine learning.
Santamouris, M., Cartalis, C., Synnefa, A., & Kolokotsa, D. (2015). On the impact of urban heat island and global warming on the power demand and electricity consumption of buildings—A review. Energy and Buildings, 98, 119-124.
Sun, Y., Haghighat, F., & Fung, B. C. (2020). A Review of the-State-of-the-Art in Data-driven Approaches for Building Energy Prediction. Energy and Buildings, 110022.
Urdan, T. C. (2016). Statistics in plain English: Taylor & Francis.
Wang, J. C. (2016). A study on the energy performance of school buildings in Taiwan. Energy and Buildings, 133, 810-822.
Wang, J. C. (2019). Analysis of energy use intensity and greenhouse gas emissions for universities in Taiwan. Journal of cleaner production, 241, 118363.
Wang, J. Q., Du, Y., & Wang, J. (2020). LSTM based long-term energy consumption prediction with periodicity. Energy, 197, 117197.
Xu, X., Wang, W., Hong, T., & Chen, J. (2019). Incorporating machine learning with building network analysis to predict multi-building energy use. Energy and Buildings, 186, 80-97.
Yang, Z., & Becerik-Gerber, B. (2014). The coupled effects of personalized occupancy profile based HVAC schedules and room reassignment on building energy use. Energy and Buildings, 78, 113-122.
Yoshino, H., Hong, T., & Nord, N. (2017). IEA EBC annex 53: Total energy use in buildings—Analysis and evaluation methods. Energy and Buildings, 152, 124-136.
Zhan, S., Liu, Z., Chong, A., & Yan, D. (2020). Building categorization revisited: A clustering-based approach to using smart meter data for building energy benchmarking. Applied Energy, 269, 114920.
Zhao, H.-x., & Magoulès, F. (2012). A review on the prediction of building energy consumption. Renewable and Sustainable Energy Reviews, 16(6), 3586-3592.
Zheng, Z., Chen, H., & Luo, X. (2019). Spatial granularity analysis on electricity consumption prediction using LSTM recurrent neural network. Energy Procedia, 158, 2713-2718.
Zhou, C., Fang, Z., Xu, X., Zhang, X., Ding, Y., & Jiang, X. (2020). Using long short-term memory networks to predict energy consumption of air-conditioning systems. Sustainable Cities and Society, 55, 102000.
Zhu, J., Shen, Y., Song, Z., Zhou, D., Zhang, Z., & Kusiak, A. (2019). Data-driven building load profiling and energy management. Sustainable Cities and Society, 49, 101587.