A Comprehensive Review of Meta-Learning Applications in Building Domain

Machine learning shows great potential in building applications but faces challenges due to system complexity and extensive data requirements. Meta-learning, an approach that enables learning from prior experiences, offers a promising solution to these challenges. Currently, studies on its applications in the building domain remain fragmented, and no comprehensive review has been conducted. Therefore, this study presents the first systematic review that synthesizes existing studies, identifies current research limitations, and outlines future opportunities. This work aims to introduce meta-learning to the building research community, providing a foundation for researchers to understand its core principles and evaluate its applicability to specific research problems.