sklearn ImportError: cannot import name plot_roc_curve

Question

I am trying to plot a Receiver Operating Characteristics (ROC) curve with cross validation, following the example provided in sklearn's documentation. However, the following import gives an ImportError, in both python2 and python3.

from sklearn.metrics import plot_roc_curve

Error:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: cannot import name plot_roc_curve

python-2.7 sklearn version: 0.20.2.

python-3.6 sklearn version: 0.21.3.

I found that the following import works fine, but it's not quite the same as plot_roc_curve.

from sklearn.metrics import roc_curve

Is plot_roc_curve deprecated? Could somebody try the code and let me know the sklearn version if it works?

Salvo

Answer 1

Plotting API was introduced in the version 0.22. As mentioned here, Scikit-learn 0.20 was the last version to support Python 2.7 and Python 3.4. Scikit-learn now requires Python 3.5 or newer.

Endothecium

Answer 2

plot_roc_curve has been removed in version 1.2. From 1.2, use RocCurveDisplay instead:

Before sklearn 1.2:

from sklearn.metrics import plot_roc_curve
svc_disp = plot_roc_curve(svc, X_test, y_test)
rfc_disp = plot_roc_curve(rfc, X_test, y_test, ax=svc_disp.ax_)

From sklearn 1.2:

from sklearn.metrics import RocCurveDisplay
svc_disp = RocCurveDisplay.from_estimator(svc, X_test, y_test)
rfc_disp = RocCurveDisplay.from_estimator(rfc, X_test, y_test, ax=svc_disp.ax_)

Kilogram

Answer 3

Install scikit-plot and import the metric from there:

from scikitplot.metrics import plot_roc_curve

Zachary

Answer 4

I updated Conda with conda update --all and then updated scikit-learn to the latest version which for me was conda install scikit-learn=0.23.2 and restarted the kernel. After that my errors were gone.

Tilney

Answer 5

As the official sklearn's document mentioned: The function plot_roc_curve is deprecated in 1.0 and will be removed in 1.2. If you would like more detail, please refer to here.

Use one of the class methods: sklearn.metric.RocCurveDisplay.from_predictions or sklearn.metric.RocCurveDisplay.from_estimator. f you would like more detail, please refer to here.

Foreclose

Answer 6

for

ImportError: cannot import name 'plot_roc_curve' from 'sklearn.metrics'

use RocCurveDisplay instead of plot_roc_curve as in from sklearn.metrics import RocCurveDisplay

Antimere

Answer 7

plot_roc_curve was deprecated and removed from sklearn in version 1.2. Use RocCurveDisplay: Instead of plot_roc_curve, the current method to plot ROC curves is through the RocCurveDisplay class in sklearn.metrics.

So try:

from sklearn.metrics import RocCurveDisplay

Squamation

Answer 8

Use RocCurveDisplay instead of plot_roc_curve:

from sklearn.metrics import RocCurveDisplay

then you can use it as such:

RocCurveDisplay.from_estimator(clf, X_test, y_test)

Coremaker