Confusion matrix predicted labels

Question

I am using the following piece of python code to generate a normalized confusion matrix

import scikitplot as skplt 
skplt.metrics.plot_confusion_matrix(y_test, y_pred, normalize=True)

and this gives me the plot as shown here. However, I wanted to rotate the texs on the x-axis (the predicted values, i.e., neutral, happy, sad). Instead of in a horizontal line, I want the predicted values on the x-axis to be rotated (as shown in here) and keep the true labels as they are (horizontal).How can we do this in Python?

Achintha Ihalage · Accepted Answer · 2020-04-17T21:14:43.363

0

If you want to use scikitplot, I believe there is x_tick_rotation argument in the plot_confusion_matrix function. So you could rotate xticks to 45 degrees as follows.

skplt.metrics.plot_confusion_matrix(y_test, y_pred, x_tick_rotation=45, normalize=True)

edited Apr 17 '20 at 21:14

answered Apr 17 '20 at 20:58

Achintha Ihalage

2,310
4
20
33

1

This is completely for a different task, sir. Please give an answer according to the simple piece of code I included. This is more confusing – Andy_Ellen Apr 17 '20 at 21:04
I have updated the answer if you want to use scikit-plot. Please see if this works – Achintha Ihalage Apr 17 '20 at 21:15
AWESOME! That's exactly what I was looking for and thank you. I mark your answer as ACCEPTED! – Andy_Ellen Apr 17 '20 at 21:18
By the way, the plot adds a "Normalized Confusion Matrix" text on its top. Is there any way to turn this text off? – Andy_Ellen Apr 17 '20 at 21:26
You can try setting `title` argument to a space (' '). `skplt.metrics.plot_confusion_matrix(y_test, y_pred, title=' ', x_tick_rotation=45, normalize=True)` – Achintha Ihalage Apr 17 '20 at 21:38
It's better to update your answer accordingly; posting code snippets in the comments should be avoided – desertnaut Apr 17 '20 at 21:59

Confusion matrix predicted labels

1 Answers1