Category Card

False discovery rate is a metric that answers the question, "How many positive predictions are incorrect?".

Macro averaging is a tool to compute a metric as a single value for multi-class problems.

Micro averaging is a tool to compute a metric as a single value for multi-class problems.

One versus all is a tool to transform a multi-class confusion matrix into several binary confusion matrices

Precision is a metric that answers the question, "How many positive predictions are correct?".

Recall - or sensitivity, or True Positive Rate - is a metric that answers the question, "How many positive example are predicted by the model?".

Specificity - or True Negative Rate - is a metric that answers the question, "How many negative examples are correctly predicted by the model?"

A Confusion Matrix is a tool to evaluate the performances of a classification model.

One versus all is a tool to transform a multi-class confusion matrix into several binary confusion matrices

Accuracy is a metric that answers the question "How many predictions are correct?".