Knowledge Maxxing← Back to topic library
ML MaxxingIntermediate75 sec
ROC curve
An ROC curve plots a classifier's true positive rate against its false positive rate across every possible classification threshold.
Prerequisite: Classification threshold
What matters
Looking at the whole curve, rather than one threshold, shows how a model trades detection for false alarms across its entire range of possible settings.
Example
A model whose ROC curve stays close to the top-left corner catches most true positives while triggering very few false positives across most thresholds.
Common misconception
An ROC curve does not tell you which single threshold to pick; it shows the range of trade-offs available, and the right choice still depends on the task's costs.
Recall question
What two rates does an ROC curve plot against each other?
Keywords and related concepts
ROC curvetrue positive ratefalse positive rateroc-auc