<p>This study investigates the thermal insulation performance of composite phase change materials (PCMs) based on a lauric acid-palmitic acid-tetradecanol ternary system, leveraging data mining techniques to identify optimal component ratios from a broad dataset prior to experimental design. Expanded perlite was used as the supporting material, with preparation via atmospheric pressure adsorption. This method addresses key challenges in applying PCMs in construction. SEM analysis confirmed stable and uniform PCM distribution within perlite pores. The composites showed excellent durability and cyclic stability, with minimal mass loss after repeated thermal cycling. Furthermore, performance data from cycling tests were integrated into a predictive model, revealing non-linear relationships between composition, pore structure, and long-term efficacy that are difficult to capture through conventional experimentation alone. These characteristics enhance indoor temperature regulation, occupant comfort, and energy efficiency in buildings. The study highlights the strong potential of PCM-based materials for sustainable construction. Future research should focus on identifying new PCMs, optimizing preparation techniques, and developing accurate simulation models to evaluate real-world energy-saving performance. These efforts could accelerate large-scale PCM adoption and support greener, energy-efficient building practices.</p>