Quantifying Performance of a Binary Classifier Through Weight of Evidence
Plot the cumulative frequency distributions in cases and in controls
Plot crude and model-based ROC curves
Plot the distribution of the weight of evidence in cases and in contro...
Proportions of cases and controls below a threshold of weight of evide...
Recalibrate posterior probabilities
Summary evaluation of predictive performance
Calculate the crude smoothed densities of W in cases and in controls
Compute densities of weights of evidence in cases and controls
Calculate weights of evidence in natural log units
Quantifying performance of a diagnostic test using the sampling distri...
Example datasets
The distributions of the weight of evidence (log Bayes factor) favouring case over noncase status in a test dataset (or test folds generated by cross-validation) can be used to quantify the performance of a diagnostic test (McKeigue (2019), <doi:10.1177/0962280218776989>). The package can be used with any test dataset on which you have observed case-control status and have computed prior and posterior probabilities of case status using a model learned on a training dataset. To quantify how the predictor will behave as a risk stratifier, the quantiles of the distributions of weight of evidence in cases and controls can be calculated and plotted.