Document recommended dataset splits

Add a compact dataset-card section describing the suggested baseline, data-efficiency, unseen-geometry, fixed-AoA, and out-of-distribution split families, with links to the manifest and detailed documentation.

README.md

@@ -49,7 +49,20 @@ Each folder (e.g., `geo_LHCi_AoA_j` where `i` is the geometry ID from 001 to 180
 In addition to the files per run folder, there are also general summary files in the root directory:
 * `geo_values_all.csv`: Reference geometry values used to define the geometry via the DoE method for all runs.
 * `force_mom_all.csv`: Time-averaged drag, lift, moment, and pressure/viscous coefficients for all runs.
-* 'splits*' : folder containing the train/val/test splits (manifest.json) and explanation and their derivation (README.pdf)
+* [`splits/`](splits/): Deterministic train/validation/test split manifest, documentation, and generation code.
+
+## Recommended Dataset Splits
+
+For reproducible machine-learning evaluation, HiLiftAeroML provides 14 deterministic split families in [`splits/manifest.json`](splits/manifest.json). Case IDs match the top-level `geo_LHC..._AoA_...` directories.
+
+| Evaluation goal | Suggested splits | Training population |
+|---|---|---:|
+| Standard baseline and case-level data efficiency | `full` → `medium` → `scarce` → `super_scarce` | 1260 → 510 → 210 → 35 cases |
+| Unseen-geometry generalization and data efficiency | `geometry` → `geometry_medium` → `geometry_scarce` → `geometry_super_scarce` | 126 → 51 → 21 → 4 complete geometries |
+| Fixed-AoA unseen-geometry evaluation | `single_aoa_4`, `single_aoa_12`, `single_aoa_22` | 126 train / 18 validation / 36 test geometries |
+| Out-of-distribution evaluation | `aoa`, `deflection`, `stall` | High-AoA, high-deflection, or post-stall extrapolation |
+
+The levels within each data-efficiency ladder are nested and share the same validation and test sets, enabling direct comparisons across training-set sizes. Use validation data for model selection and reserve test data for final evaluation. See [`splits/README.md`](splits/README.md) for the complete definitions, counts, and selection guidance, or [`splits/generate_splits.py`](splits/generate_splits.py) for the generation and validation logic.
 
 ## How to Download
 
@@ -174,4 +187,4 @@ This dataset is provided under the permissive open-source **CC-BY-4.0** license.
 
 * 26/05/2026 - IDs 0-43 for angles 16-22 had incomplete force_mom (didn't have std dev etc). Corrected. Please download forces again if you took files before this date
 
-* 20/05/2026 - dataset fully uploaded and arxiv preprint made available
+* 20/05/2026 - dataset fully uploaded and arxiv preprint made available