Expected Information Gain Estimation on Sign-Ambiguous Bimodal Model d=32, |\u03be|=8, seed 42
21.09Expected Information Gain (EIG)Exact
Evaluation results
| Method | Links | |
|---|---|---|
| ExactDesign=A, Composition=8 strong2026.04 | 21.09 | |
| ExactDesign=B, Composition=8 medium2026.04 | 15.53 | |
| ExactDesign=D, Composition=4 strong + 4 null2026.04 | 11.61 | |
| ExactDesign=C, Composition=8 weak2026.04 | 10 | |
| ExactDesign=E, Composition=8 null2026.04 | 2.12 | |
| Gaussian VBDesign=E, Composition=8 null2026.04 | 1.23 | |
| Gaussian VBDesign=C, Composition=8 weak2026.04 | -2.53 | |
| Gaussian VBDesign=D, Composition=4 strong + 4 null2026.04 | -6.34 | |
| Gaussian VBDesign=B, Composition=8 medium2026.04 | -8.27 | |
| Gaussian VBDesign=A, Composition=8 strong2026.04 | -13.91 |