yardstick 0.0.5

Other improvements

• The autoplot() heat map for confusion matrices now places the predicted values on the x axis and the truth values on the y axis to be more consistent with the confusion matrix print() method.

• The autoplot() mosaic plot for confusion matrices had the x and y axis labels backwards. This has been corrected.

yardstick 0.0.4

New metrics and functionality

• iic() is a new numeric metric for computing the index of ideality of correlation. It can be seen as a potential alternative to the traditional correlation coefficient, and has been used in QSAR models (@jyuu, #115).

• average_precision() is a new probability metric that can be used as an alternative to pr_auc(). It has the benefit of avoiding any issues of ambiguity in the case where recall == 0 and the current number of false positives is 0.

Other improvements

• metric_set() output now includes a metrics attribute which contains a list of the original metric functions used to generate the metric set.

• Each metric function now has a direction attribute attached to it, specifying whether to minimize or maximize the metric.

• Classification metrics that can potentially have a 0 value denominator now throw an informative warning when this case occurs. These include recall(), precision(), sens(), and spec() (#98).

• The autoplot() method for pr_curve() has been improved to always set the axis limits to c(0, 1).

• All valid arguments to pROC::roc() are now utilized, including those passed on to pROC::auc().

• Documentation for class probability metrics has been improved with more informative examples (@rudeboybert, #100).

Bug fixes

• mn_log_loss() now uses the min/max rule before computing the log of the estimated probabilities to avoid problematic undefined log values (#103).

• pr_curve() now places a 1 as the first precision value, rather than NA. While NA is technically correct as precision is undefined here, 1 is practically more correct because it generates a correct PR Curve graph and, more importantly, allows pr_auc() to compute the correct AUC.

• pr_curve() could generate the wrong results in the somewhat rare case when two class probability estimates were the same, but had different truth values.

• pr_curve() (and subsequently pr_auc()) now generates the correct curve when there are duplicate class probability values (reported by @dariyasydykova, #93).

• Binary mcc() now avoids integer overflow when the confusion matrix elements are large (#108).

v0.0.3

• A few new metrics from tidyverse developer day.
• Fixed a few yardstick related bugs.
• Updated tests to comply with the R 3.6 sample() fiasco.

v0.0.2

• Breaking changes regarding stabilizing the API
• Multiclass support
• Curve metrics
• More classification and regression metrics
• Altered documentation to be 1 help file per metric