Supporting Information for Interpretable Machine Learning Modeling of Bulk Modulus in Cubic Perovskites and Its Application to Soft-Lattice Screening

The supporting information includes four supplementary tables: Table S1 summarizes 22 elemental features generated by the Magpie tool, including English names, symbols, and Chinese names; Table S2 provides the elemental feature values of 213 perovskite samples used in the SISSO model, the bulk moduli predicted based on the thermal-structural coupling descriptors, and the corresponding DFT calculated values; Table S3 corresponds to the VS-SISSO model, listing the elemental feature values, DFT moduli, and the bulk moduli predicted using the electron-thermal-structural triple coupling descriptors for the same set of perovskite samples; Table S4 compiles 169 lead-free perovskite candidate materials with bulk moduli similar to those of Pb-I perovskites and their predicted values, which were screened out through this descriptor.