Exam Score Prediction 10-fold OOF 1.0 (held-out)
8.582RMSEFull Regularized Ensemble
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Full Regularized Ensemblecomponents=Ridge+Lasso+ElasticNet Blend, # Models=372026.02 | 8.582 | — | 6.546 | 0.7872 | 0.887 | -0.021 | 0.639 | |
| Ridge Meta-Learnerlambda=0.87, # Models=372026.02 | 8.583 | — | 6.547 | 0.7871 | 0.8869 | -0.02 | 0.682 | |
| Lasso Meta-Learnerlambda=0.03, sparsity=68%, # Models=372026.02 | 8.584 | — | 6.548 | 0.7871 | 0.8868 | -0.019 | 0.695 | |
| ElasticNet Meta-Learnerlambda1=0.05, alpha=0.5, # Models=372026.02 | 8.584 | — | 6.548 | 0.7871 | 0.8868 | -0.019 | 0.695 | |
| Hill Climbing (Greedy)strategy=Greedy, # Models=282026.02 | 8.603 | — | 6.561 | 0.7861 | 0.8863 | — | — | |
| Vanilla Stacking (Ridge)lambda=1.2, # Models=452026.02 | 8.627 | — | 6.578 | 0.7848 | 0.8856 | 0.045 | 0.012 | |
| Vanilla Stacking (Linear)mode=Linear Stacking, # Models=452026.02 | 8.691 | — | 6.634 | 0.7815 | 0.8841 | 0.109 | — | |
| Weighted Average (Performance)# Models=452026.02 | 8.756 | — | 6.691 | 0.7782 | 0.8826 | 0.174 | — | |
| Simple Average (All)# Models=452026.02 | 8.894 | — | 6.812 | 0.771 | 0.8781 | 0.312 | — | |
| Best Single Model# Models=12026.02 | 9.247 | — | 7.103 | 0.7521 | 0.8674 | 0.665 | — |