How the model was chosen
Five candidate models and a baseline that always predicts the mean took the same test: the data was shuffled and split into 10 parts, each part was held out once for testing, and this was repeated 5 times. The table shows the mean and standard deviation over these 50 tests.
| Model | Mean error (g) | RMSE (g) | R² |
|---|
Why the simplest model
Once head size is known, sex and age group add almost no new information. Random Forest overfits on such a small, linear dataset and did worse.