Skip to content

Classification performance standards such as accuracy, precision, recall, confusion matrix, F1, ROC, AUC. Take MNIST as example.

License

Notifications You must be signed in to change notification settings

vba34520/Classification-Performance-Standard

Repository files navigation

Classification performance standard

Description

001 Classification-Performance-Standard.ipynb Classification performance standards such as accuracy, precision, recall, confusion matrix, F1, ROC, AUC. Take MNIST as example.

002 Classification-Performance-Standard.ipynb The framework to evaluate multi-classification classifier automatically.

Standard

Confusion matrix

Predict Positive Predict Negative
Actual Positive TP(True Positive) FN(False Negative)
Actual Negative FP(False Positive) TN(True Negative)

Confusion matrix

The vertical axis of the ROC(Receiver Operating Characteristic) curve is TPR(True Positive Rate) and the horizontal axis is FPR(False Positive Rate).

AUC is the area under ROC curve.

ROC

ROC(each class)

About

Classification performance standards such as accuracy, precision, recall, confusion matrix, F1, ROC, AUC. Take MNIST as example.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published