Meta-learning efficiency evaluation on Accumulated N=630,000 (val)
3.2Time (s)Simple Average
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Simple AverageCPU core=single, Initial base models (K)=452026.02 | 3.2 | 0.3 | 124 | |
| Weighted AverageCPU core=single, Initial base models (K)=452026.02 | 12.7 | 1.3 | 128 | |
| Vanilla Stack (Linear)CPU core=single, Initial base models (K)=452026.02 | 67.4 | 6.7 | 245 | |
| Vanilla Stack (Ridge)CPU core=single, Initial base models (K)=452026.02 | 189.3 | 18.9 | 251 | |
| Ridge OnlyCPU core=single, Initial base models (K)=452026.02 | 234.7 | 23.5 | 268 | |
| Lasso OnlyCPU core=single, Initial base models (K)=452026.02 | 287.3 | 28.7 | 271 | |
| ElasticNet OnlyCPU core=single, Initial base models (K)=452026.02 | 301.2 | 30.1 | 273 | |
| Full PipelineCPU core=single, Initial base models (K)=45, Components=redundancy projection + regularization2026.02 | 712.8 | 71.3 | 289 | |
| Hill ClimbingCPU core=single, Initial base models (K)=452026.02 | 2,841.6 | 284.2 | 387 |