Receiver operating characteristicsIt is a test indicator of the system matching algorithm. It is a relationship between the matching score threshold, false recognition rate and rejection rate. It reflects the balance between the rejection rate and false recognition rate of the recognition algorithm at different thresholds.
True category
Prediction is positive
Prediction is negative
Positive Example
TP (True Positive)
FN (False Negative Example)
Counterexample
FP (False Positive)
TN (True Counterexample)
The ROC curve is a curve graph with the false positive rate FPR as the horizontal axis and the true positive rate TPR as the vertical axis, and is defined as follows:
TPR: The ratio of samples that are correctly judged as positive among all samples that are actually positive. TPR = TP / (TP + FN)
FPR: The ratio of samples that are falsely judged as positive among all samples that are actually negative. FPR = FP / (FP + TN)
The ROC curve can be used to calculate the Mean Average Precision, which is the average accuracy obtained by changing the threshold to select the best result. Generally speaking, the closer the curve is to the upper left corner, the better the classifier effect.
