1
$\begingroup$

How do I add class labels to the confusion matrix? The label display number in the label not the actual value of the label Eg. labels = ['A','B','C','D','E','F','G','H','I','J','K','L','M','N','O','P','Q','R','S','T','U','V','W','X','Y','Z']

confusion matrix image

Here is the code I used to generate it.

x_train, y_train, x_test, y_test = train_images, train_labels, test_images, test_labels
model = KNeighborsClassifier(n_neighbors=7, metric='euclidean')
model.fit(x_train, y_train)
# predict labels for test data
predictions = model.predict(x_test)

# Print overall accuracy
print("KNN Accuracy = ", metrics.accuracy_score(y_test, predictions))

# Print confusion matrix
cm = confusion_matrix(y_test, predictions)

plt.subplots(figsize=(30, 30))         # Sample figsize in inches
sns.set(font_scale=2)
sns.heatmap(cm, annot=True)
$\endgroup$

3 Answers 3

1
$\begingroup$

You can use the xticklabels and yticklabels arguments for this as specified in the seaborn documentation. See also the following answer on a different question on the datascience stackexchange.

$\endgroup$
0
1
$\begingroup$

You can also pass the correctly labeled predictions to the confusion matrix.

from sklearn.preprocessing import LabelEncoder
from sklearn.metrics import ConfusionMatrixDisplay

labelEnc = LabelEncoder()
y = labelEnc.fit_transform(...) # Encode string-labels with numbers.
ConfusionMatrixDisplay.from_predictions(
    labelEnc.inverse_transform(y_test), # Decode numbers to string-labels.
    labelEnc.inverse_transform(y_pred))
```
$\endgroup$
1
$\begingroup$

You can first define your labels an use that in xticklabels and yticklabels values.

import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.metrics import confusion_matrix
from sklearn.neighbors import KNeighborsClassifier
from sklearn import metrics

#defined your labels
labels = ['A','B','C','D','E','F','G','H','I','J','K','L','M','N','O','P','Q','R','S','T','U','V','W','X','Y','Z']

# you have already trained your model and obtained predictions

# Print confusion matrix
cm = confusion_matrix(y_test, predictions)

# Plot confusion matrix with class labels
plt.subplots(figsize=(30, 30))
sns.set(font_scale=2)
sns.heatmap(cm, annot=True, xticklabels=labels, yticklabels=labels)
plt.xlabel('Predicted')
plt.ylabel('True')
plt.show()
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.