Compact and Physically Interpretable Feature Models for Photometric Type Ia Supernova Classification

Abstract

Photometric classification of Type Ia supernovae is essential for modern time-domain surveys, where spectroscopic confirmation is not always feasible for the full transient sample. We investigate a compact and physically interpretable feature representation derived from multi-band light curves and evaluate its performance using gradient-boosted decision trees on the Supernova Photometric Classification Challenge (SPCC) dataset. The compact representation is derived from an initial pool of 31 light-curve features, reduced to 30 after removing redundant variables and further optimized to a 16-feature model through systematic ablation analysis. The final compact model achieves an F1-score of 0.844 on the held-out test set, consistent with k-fold cross-validation results (0.841 +/- 0.006). The precision-recall area under the curve (PR-AUC) is 0.928, with similarly low variance across folds. Ablation experiments show that temporal evolution provides the dominant classification signal, while brightness, color, and variability features contribute complementary information. A reduced core of approximately ten physically meaningful features retains most of the performance of the compact model, with only a small decrease in F1-score, indicating that reliable classification does not require large high-dimensional feature spaces. These results demonstrate that interpretable feature-based models can capture the essential astrophysical information needed for Type Ia photometric classification, with implications for survey cadence, filter coverage, and the design of transparent and efficient machine-learning pipelines for time-domain surveys.