List your company
ResearchFormerly WizWand

Linear Regression on Simulated data

0.156Calibration ErrorSandwich Gauss
0.141080.241790.34250.44321May 29, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Sandwich GaussB=162026.050.1560
Sandwich GaussB=⌊0.1 × N⌋2026.050.1560
DQ+exactB=16, Loss=Log loss2026.050.1690.664
CTB=16, Loss=Log loss2026.050.1710.672
DQ+exactB=16, Loss=β-loss (β = 1.5)2026.050.1720.695
DQ+exactB=⌊0.1 × N⌋, Loss=Log loss2026.050.1740.672
LR+WSB=⌊0.1 × N⌋, Loss=β-loss (β = 1.5)2026.050.1771.322
LR+WSB=16, Loss=β-loss (β = 1.5)2026.050.1781.115
CTB=⌊0.1 × N⌋, Loss=Log loss2026.050.1790.975
DQ+exactB=⌊0.1 × N⌋, Loss=β-loss (β = 1.5)2026.050.190.748
NUTSB=162026.050.1950.795
NUTSB=⌊0.1 × N⌋2026.050.1950.799
CTB=⌊0.1 × N⌋, Loss=β-loss (β = 1.5)2026.050.1961.006
CTB=16, Loss=β-loss (β = 1.5)2026.050.2010.64
PosteriorB=16, Loss=Log loss2026.050.4180.943
PosteriorB=⌊0.1 × N⌋, Loss=Log loss2026.050.4180.943
LR+WSB=⌊0.1 × N⌋, Loss=Log loss2026.050.5170.996
LR+WSB=16, Loss=Log loss2026.050.5290.995