Free research tools · Page 8

Binary log loss by forecast

Find the average log penalty and detect impossible forecasts for observed outcomes.

Identify the weakest class recall and compare macro recall with overall accuracy.

Compare a fixed decision rule with treating everyone or no one at a stated threshold.

Rank likelihood fits and inspect their relative support within a candidate set.

Compute the unpooled standard error and Welch–Satterthwaite degrees of freedom.

Compare crude and directly standardized rates using explicit reference weights.

Apply reference rates to stratum-specific exposure and audit an indirectly standardized ratio.

Combine stratum cross-products using Mantel–Haenszel weighting.

Estimate the chance of finding at least one defect in a random audit without replacement.

Bound an event probability after zero events using an exact one-sided binomial limit.

Find the probability that a lot passes a maximum-defect acceptance rule.

Quantify the residual variation explained by adding predictors to a nested linear model.