mmetrics is a grammar of metrics calculation, providing a consistent set of functions that help you quantitatively assess the performance of statistical models:
-
confusion_scores
andmulticlass_confusion_scores
provides confusion matrices given model predictions and ground truth labels. -
accuracy function (
binary_acc
andmulticlass_acc
) estimates the correctness of predicted classification labels. -
precision function (
binary_precision
andmulticlass_precision
) measures the accuracy of positive predictions. -
recall function (
binary_recall
andmulticlass_recall
) measures the completeness of positive predictions. -
f1 score computation (
binary_f1
andmulticlass_f1
) gives a balanced view of precision and recall scores, reflecting how well a classification model is to assign each observation to their corresponding calss.
install.packages("mmetrics")
Please refer to doc/Tutorial.html
for example usage or consult function help page.