Machine learning application in building energy consumption prediction: A comprehensive review

Energy, the lifeblood of modern society, has garnered increased attention toward managing building energy consumption over the past decade. The proliferation of available data has opened new avenues for leveraging machine learning to predict building energy usage. Despite the extensive literature in this domain, there is a lack of systematic reviews that provide a comprehensive overview of machine learning applications in predicting building energy consumption. This research aims to: (1) review the diverse applications of machine learning in forecasting building energy consumption; (2) summarize recent advancements in machine learning for enhancing building energy efficiency; and (3) identify current research gaps while proposing future trends. Initially, 431 relevant articles published between 2012 and 2023 were examined using bibliometric analysis, leading to the identification of 16 research keywords and 9 clusters. Subsequently, content analysis was employed to assess building types, energy sources, and temporal granularity. Finally, existing research gaps were identified, and six future research directions were proposed, including (1) integration of multi-source heterogeneous data; (2) development of model transfer frameworks across different buildings; (3) increased focus within the computer science community; (4) interdisciplinary collaboration and standardization; (5) real-time monitoring and forecasting of energy consumption; and (6) data security and privacy protection. This research not only highlights prevailing research gaps but also outlines future trajectories, providing valuable insights for researchers and practitioners navigating this dynamic field.