Online Resource 1: Supplementary Material for "Reliable Machine Learning for Small Grouped Geotechnical–Geophysical Datasets: Dependency-Aware Feature Engineering and Nested Group Validation"

This record contains Online Resource 1, the supplementary material accompanying the manuscript “Reliable Machine Learning for Small Grouped Geotechnical–Geophysical Datasets: Dependency-Aware Feature Engineering and Nested Group Validation.” The supplementary material provides detailed documentation and diagnostic evidence supporting the reported machine-learning analysis. It includes: (1) the complete engineered-feature catalogue and exact feature definitions; (2) the complete membership of the seven candidate feature sets used in the analysis; (3) sensitivity results for the exclusion of flagged observations; (4) five-seed sensitivity results for the locked model configurations; (5) fold-level performance and sample/group counts for the primary and secondary models; (6) complete five-seed fold-level sensitivity results; and (7) diagnostic figures showing observed-versus-predicted values and residuals for true resistivity and S-wave velocity. The material is provided to support methodological transparency, result interpretation, and computational reproducibility of the study. This record contains supplementary documentation and diagnostic outputs and does not constitute the original measurement-level dataset used to generate the reported results.