Supporting Dataset for the mRMR–GPR Analysis of XRD-Derived Predictors of Perovskite Film Stability

This dataset contains the X-ray diffraction (XRD)-derived descriptors and target values used for the hierarchical minimum-redundancy maximum-relevance–Gaussian process regression (mRMR–GPR) analysis reported in the associated study. The modelling table contains 24 candidate XRD descriptors and one target variable. The descriptors include peak-area, full-width-at-half-maximum, and texture-related parameters extracted from the XRD patterns of perovskite thin-film samples prepared under different processing conditions.The model input comprised 24 XRD-derived descriptors. Features B–I correspond to the peak areas, PA, of the (100), (110), (111), (200), (210), (211), (220), and (221) reflections, respectively. Features J–Q represent the corresponding full widths at half maximum, FWHM, for the same sequence of reflections, while features R–Y represent the corresponding texture coefficients, TC. Accordingly, the three descriptors retained in the final model were I = PA(221), P = FWHM(220), and D = PA(111).