Confusion Matrix

Binary Classification

import matplotlib.pyplot as plt

from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay

import numpy as np

# Example data for a binary classification problem

y_true = np.array([0, 1, 0, 1, 1, 0])

y_pred = np.array([0, 1, 1, 1, 0, 0])

# 1. Create the confusion matrix

cm = confusion_matrix(y_true, y_pred)

# 2. Instantiate the ConfusionMatrixDisplay

disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=['Class 0', 'Class 1'])

#3  Calculate True Positives, True Negatives, False Positive and False Negatives 

TN, FP, FN, TP = cm.ravel()  # for binary classification

print(f' TP = {TP}, TN = {TN}, FP = {FP}, FN = {FN}')

# 4. Plot the confusion matrix and get the axes object

disp.plot(cmap=plt.cm.Blues) 

Note : In binary classification, recall of the positive class is also known as "sensitivity"; recall of the negative class is "specificity

Multi Class Classification

import matplotlib.pyplot as plt

from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay

import numpy as np

# Example data for a binary classification problem

y_true1 = [0, 1, 2, 2, 2]

y_pred1 = [0, 0, 2, 2, 1]

# 1. Create the confusion matrix

cm1 = confusion_matrix(y_true1, y_pred1)

print(type(cm1))

print(cm1)

# 2. Instantiate the ConfusionMatrixDisplay

target_names = ['class 0', 'class 1', 'class 2']

disp1 = ConfusionMatrixDisplay(confusion_matrix=cm1, display_labels=target_names)

3  Calculate True Positives, True Negatives, False Positive and False Negatives 

num_classes = cm1.shape[0]  # Get the number of classes

TP = np.diag(cm1)  # True Positives for each class are on the diagonal

FP = cm1.sum(axis=0) - TP  # False Positives are the sum of the column minus the TP

FN = cm1.sum(axis=1) - TP  # False Negatives are the sum of the row minus the TP

TN = cm1.sum() - (TP + FP + FN) # True Negatives for each class are all other correct predictions

print("True Positives per class:", TP)

print("True Negatives per class:", TN)

print("False Positives per class:", FP)

print("False Negatives per class:", FN)

# 4. Plot the confusion matrix and get the axes object