Predicting open-plan office window operating behavior using the random forest algorithm

Abstract Understanding window operating behavior in offices is important in terms of its influence on reducing energy consumption and improving indoor comfort. Researchers have applied different mathematical methods to develop useful window operating behavior models, however, the applicable machine learning algorithms are still in their preliminary research stage, requiring additional development. In the work described here, the authors applied the random forest (RF) algorithm to predict window operating behavior in open-plan offices, using data from three such offices in Nanjing, Jiangsu Province, China. The three open-plan offices were different in terms of their areas, office types, numbers of occupants, and layouts. The importance of various elements influencing window operating behavior was determined through the RF method, and the resulting rankings were consistent with the occupants' subjective understanding, as determined using a questionnaire. The sensitivity of the RF model to the number of inputs was explored, and showed that, with four inputs, its accuracy could reach 80%. The RF model also showed high accuracy and stability in predicting window operating behavior in two application formats—namely, for different offices, and for the same office over different years. Meanwhile, the RF model was compared with the other two popular machine learning methods, namely SVM and XGBoost algorithms, which also proves the high accuracy of RF models. The results obtained in this study should provide insights into applying machine learning methods to window operating behavioral studies generally—and may also inspire their application to other behavioral study types.

[1]  Shen Wei,et al.  Developing window behavior models for residential buildings using XGBoost algorithm , 2019, Energy and Buildings.

[2]  M. El Mankibi,et al.  Occupant presence and behavior: A major issue for building energy performance simulation and assessment , 2020 .

[3]  David M. Schnyer,et al.  Support vector machine , 2020, Machine Learning.

[4]  Stefano Paolo Corgnati,et al.  Verification of stochastic behavioural models of occupants' interactions with windows in residential buildings , 2015 .

[5]  Hom B. Rijal,et al.  Development of a window opening algorithm based on adaptive thermal comfort to predict occupant behavior in Japanese dwellings , 2018, Japan Architectural Review.

[6]  S. Sharples,et al.  Thermal comfort, occupant control behaviour and performance gap – A study of office buildings in north-east China using data mining , 2019, Building and Environment.

[7]  Dirk Van den Poel,et al.  Predicting customer retention and profitability by using random forests and regression forests techniques , 2005, Expert Syst. Appl..

[8]  Pedro F. Pereira,et al.  Data-driven occupant actions prediction to achieve an intelligent building , 2019, Building Research & Information.

[9]  Mohamed El Mankibi,et al.  Analysis on the driving factors and patterns of window opening and closing behaviour in French households , 2019, IOP Conference Series: Materials Science and Engineering.

[10]  Xin Zhou,et al.  Comparative analysis of window operating behavior in three different open-plan offices in Nanjing , 2021 .

[11]  B. Dong,et al.  A survey on energy consumption and energy usage behavior of households and residential building in urban China , 2017 .

[12]  Elie Azar,et al.  A systematic review of occupant behavior in building energy policy , 2020 .

[13]  Jin Wen,et al.  Simulating the human-building interaction: Development and validation of an agent-based model of office occupant behaviors , 2015 .

[14]  Tianzhen Hong,et al.  State-of-the-art on research and applications of machine learning in the building life cycle , 2020, Energy and Buildings.

[15]  I. Hung,et al.  Carbon dioxide concentration as indicator of indoor air quality , 1989 .

[16]  Yanjun Qi.  Random Forest for Bioinformatics , 2012 .

[17]  Francesca Stazi,et al.  Modelling window status in school classrooms. Results from a case study in Italy , 2017 .

[18]  Tom Fawcett,et al.  An introduction to ROC analysis , 2006, Pattern Recognit. Lett..

[19]  K. Steemers,et al.  Night-time naturally ventilated offices: Statistical simulations of window-use patterns from field monitoring , 2010 .

[20]  Taehoon Hong,et al.  Automatic ventilation control algorithm considering the indoor environmental quality factors and occupant ventilation behavior using a logistic regression model , 2019, Building and Environment.

[21]  Andreas Wagner,et al.  Does the occupant behavior match the energy concept of the building? - Analysis of a German naturally ventilated office building , 2015 .

[22]  Junseok Park,et al.  Machine learning algorithms for predicting occupants' behaviour in the manual control of windows for cross-ventilation in homes , 2020, Indoor and Built Environment.

[23]  Darren Robinson,et al.  Interactions with window openings by office occupants , 2009 .

[24]  Xuyang Zhong,et al.  Verification of behavioural models of window opening: The accuracy of window-use pattern, indoor temperature and indoor PM2.5 concentration prediction , 2020 .

[25]  Yixuan Wei,et al.  Comparison of different window behavior modeling approaches during transition season in Beijing, China , 2019, Building and Environment.

[26]  Vikas Chandan,et al.  Non-intrusive estimation and prediction of residential AC energy consumption , 2016, 2016 IEEE International Conference on Pervasive Computing and Communications (PerCom).

[27]  Hiroshi Yoshino,et al.  IEA EBC annex 53: Total energy use in buildings—Analysis and evaluation methods , 2017 .

[28]  Verena Marie Barthelmes,et al.  Exploration of the Bayesian Network framework for modelling window control behaviour , 2017 .

[29]  D. Yan,et al.  Modelling of energy consumption and carbon emission from the building construction sector in China, a process-based LCA approach , 2019, Energy Policy.

[30]  Tianqi Chen,et al.  XGBoost: A Scalable Tree Boosting System , 2016, KDD.

[31]  Philip H. S. Torr,et al.  Randomized trees for human pose detection , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.

[32]  Christoph van Treeck,et al.  Window Opening Model using Deep Learning Methods , 2018, Building and Environment.

[33]  Donglin Zeng,et al.  Reinforcement Learning Trees , 2015, Journal of the American Statistical Association.

[34]  Bjarne W. Olesen,et al.  Window opening behaviour modelled from measurements in Danish dwellings , 2013 .

[35]  Fu Xiao,et al.  A short-term building cooling load prediction method using deep learning algorithms , 2017 .

[36]  David A. Landgrebe,et al.  A survey of decision tree classifier methodology , 1991, IEEE Trans. Syst. Man Cybern..

[37]  D. F. Parkhurst,et al.  Indicator bacteria at five swimming beaches-analysis using random forests. , 2005, Water research.

[38]  Wanyue Chen,et al.  Research on occupants’ window opening behavior in residential buildings based on the survival model , 2020 .

[39]  Yuan Jin,et al.  A novel reinforcement learning method for improving occupant comfort via window opening and closing , 2020, Sustainable Cities and Society.

[40]  Alberto Giretti,et al.  Stochastic behavioural models of occupants' main bedroom window operation for UK residential buildings , 2017 .

[41]  Gerardo Maria Mauro,et al.  Is it fundamental to model the inter-building effect for reliable building energy simulations? Interaction with shading systems , 2020 .

[42]  Peter V. Nielsen,et al.  Seasonal variation of window opening behaviors in two naturally ventilated hospital wards , 2017, Building and Environment.

[43]  Christoph van Treeck,et al.  Comparison of Different Classification Algorithms for the Detection of User's Interaction with Windows in Office Buildings , 2017 .

[44]  Halil Zafer Alibaba,et al.  Window Design of Naturally Ventilated Offices in the Mediterranean Climate in Terms of CO2 and Thermal Comfort Performance , 2020 .

[45]  Yuan Jin,et al.  A data-driven model predictive control for lighting system based on historical occupancy in an office building: Methodology development , 2020, Building Simulation.

[46]  Xin Zhou,et al.  An action-based Markov chain modeling approach for predicting the window operating behavior in office spaces , 2020, Building Simulation.