mranuraggarg/supernovae_classification: Phase 2 Tier 2 v2: Compact Feature Ablation Support Package

Phase 2 Tier 2 v2 release for the canonical compact-feature ablation branch.\n\nThis version supersedes the initial phase2-tier2-v1.0 archive and reflects the normalized Phase 2 Tier 2 branch state.\n\nHighlights:\n- Standardizes former revision support material under Phase 2 Tier 2 naming.\n- Adds dataset, feature-dictionary, training-protocol, selected-model, uncertainty, learning-curve, example-light-curve, audit, and artifact-index support files under results/phase2_tier2.\n- Updates README.md to describe the current Phase 2 Tier 2 branch state and release/DOI guidance.\n- Keeps the frozen compact 16-feature baseline reference: F1 0.844230, ROC-AUC 0.976588, PR-AUC 0.927761.\n- Preserves the corrected Phase 2 Tier 1 compact-baseline ROC and precision-recall curves for the 16-feature model.\n- Retains Tier 2 ablation, block-ablation, subset-growth, and minimal-core outputs and figures.\n\nPrevious Phase 2 Tier 2 v1 DOI: https://doi.org/10.5281/zenodo.19665796\nTarget branch: origin/phase2-tier2-compact-feature-ablation\nTarget commit: 86fc5daac4a6d85e39c804cfb8ca15afe77a928a