错误绘制来自 Sklearn 的混淆矩阵

Question

I am trying to use the sklearn confusion matrix class to plot a confusion matrix.

Here's the code I used:

from sklearn.metrics import roc_curve, auc, plot_confusion_matrix

import matplotlib.pyplot as plt

disp = plot_confusion_matrix(self.g_cv.best_estimator_ , self.test_X, self.test_Y,
                                 cmap=plt.cm.Blues)

        plt.title('Confusion Matrix')
        plt.plot(disp)

This is based on this example on Sklearn website:

https://scikit-learn.org/stable/modules/generated/sklearn.metrics.plot_confusion_matrix.html

However I get an error when this is called.

TypeError: float() argument must be a string or a number, not 'ConfusionMatrixDisplay'

It does plot it but also throws this error. Is there something wrong in the return object which is being plotted?

Answer 1

The error is raised when trying to plot the ConfusionMatrixDisplay-object that disp = plot_confusion_matrix(... returns. It is meant to create a plot right away.

It is sufficient to use plt.show() , rather than plt.plot(disp) . matplotlib.pyplot.plot() expects two arrays as input.

Either you can plot the confusion matrix right away with

disp = plot_confusion_matrix(self.g_cv.best_estimator_ , self.test_X, self.test_Y, cmap=plt.cm.Blues)
plt.show()

or you create it first and then display it

disp = ConfusionMatrixDisplay(self.g_cv.best_estimator_ , self.test_X, self.test_Y)
disp = disp.plot(cmap=plt.cm.Blues)
plt.show()