A data-driven approach is proposed to accurately predict individualized thermal comfort conditions by integrating the ASHRAE Global Thermal Comfort Database II with subject-based thermal comfort profiles of building occupants. By applying transfer learning techniques, this framework generates subject-specific models that enhance accuracy despite limited data points. Combinatorial probability analysis indicates that achieving total satisfaction becomes increasingly challenging as the group size increases. The limitations of current standards expose the pressing need for granular zoning strategies, with a data-driven, personalized thermal comfort modeling framework offering a more responsive, efficient, and personalized approach to climate control in shared spaces.
