Research and Applications of Data Mining Techniques for Improving Building Operational Performance
Purpose of Review
This paper reviews the data mining (DM)-related research and applications at the building operation stage. It aims to summarize DM-based solutions for building energy management and reveal current research and development outcomes in analyzing massive building operational data using advanced DM techniques.
Previous studies mainly adopt DM techniques for two tasks, i.e., (1) predictive modeling; (2) fault detection and diagnosis. The knowledge discovered has been successfully utilized to facilitate the decision-making during building operations. Domain expertise play the dominant role in the knowledge discovery process, which limits the chance of discovering novel knowledge.
DM is a promising technology for the development of intelligent and automated building management systems. Despite encouraging results, more research efforts should be made in (1) exploring the usefulness of unsupervised DM, (2) developing generic analytic frameworks, and (3) analyzing unstructured and multi-relational data sets.
KeywordsBig data Data mining Knowledge discovery Building operational performance Building energy management Intelligent building
The authors gratefully acknowledge the support of this research by the Research Grant Council of the Hong Kong SAR (152181/14E) and the Natural Science Foundation of SZU (Grant No. 2017061).
Compliance with Ethical Standards
Conflict of Interest
The authors declare that they have no conflict of interest.
Human and Animal Rights and Informed Consent
This article does not contain any studies with human or animal subjects performed by any of the authors.
Papers of particular interest, published recently, have been highlighted as: • Of importance •• Of major importance
- 2.Waide P, Ure J, Karagianni N, Smith G, Bordass B. The scope for energy and CO2 savings in the EU through the use of building automation technology. Final Report for the European Copper Institute, August 2013.Google Scholar
- 3.Gantz J, Reinsel D. The digital universe in 2020: big data, bigger digital shadows, and biggest growth in the far east. International Data Corporation, IDC iView: IDC Analyze the Future, 2012.Google Scholar
- 4.Han JW, Kamber M. Data mining: concepts and techniques. The Morgan Kaufmann Series in Data Management Systems; 2011.Google Scholar
- 7.•• Fan C, Xiao F, Li ZD, Wang JY. Unsupervised data analytics in mining big building operational data for energy efficiency enhancement: a review. Energy Build. 2018;159:296–308. The paper provides a comprehensive review on the use of unsupervised data analytics in analyzing big building operational data. CrossRefGoogle Scholar
- 9.• Wei YX, Zhang XX, Shi Y, Xia L, et al. A review of data-driven approaches for prediction and classification of building energy consumption. Renew Sust Energ Rev. 2018;82:1027–47. The paper serves as an updated review on the status-quo of data-driven techniques for building energy consumption. CrossRefGoogle Scholar
- 32.• Fan C, Xiao F, Zhao Y. A short-term building cooling load prediction method using deep learning algorithms. Appl Energy. 2017;195:222–33. The paper investigates the performance of deep learning in predicting building cooling load. It validates the power of unsupervised deep learning in deriving useful high-level input variables. CrossRefGoogle Scholar
- 57.• Fan C, Xiao F, Zhao Y. Wang JY. Analytical investigation of autoencoder-based methods for unsupervised anomaly detection in building energy data. Appl Eenrgy. 2018;211:1123–35. The paper investigates the power of different autoencoders in detecting anomalies in building energy data in an unsupervised way. CrossRefGoogle Scholar
- 60.•• Miller C, Nagy Z, Schlueter A. A review of unsupervised statistical learning and visual analytics techniques applied to performance analysis of non-residential buildings. Renew Sust Energy Rev. 2017;81:1365–77. The paper summarizes the applications of unsupervised data analytics and visualization techniques in analyzing building data. CrossRefGoogle Scholar