confusion_matrix: Confusion Matrix Metric

The confusion_matrix function computes the confusion matrix for classification tasks. This metric provides a summary of prediction results, showing how many samples were correctly or incorrectly classified for each class.

---

Overview

A confusion matrix is a table that is often used to describe the performance of a classification model.
- Each row represents the actual class. - Each column represents the predicted class. - Entry (i, j) is the number of samples with true label i and predicted label j.

This metric helps you understand the types of errors your classifier is making and is essential for evaluating classification accuracy, precision, recall, and other related metrics.

• Compatibility: Accepts NumPy arrays and Pandas Series for y_true and y_pred.

---

Parameters

Parameter	Type	Description	Default
y_true	array-like or pandas Series	Ground truth (correct) target values. Shape: (n_samples,)	Required
y_pred	array-like or pandas Series	Estimated targets as returned by a classifier. Shape: (n_samples,)	Required
labels	array-like	List of labels to index the matrix. If None, uses all labels in sorted order.	None
verbose	bool	If True, enables detailed logging for debugging.	False

---

Returns

• ndarray
  Confusion matrix of shape (n_classes, n_classes). Entry (i, j) is the number of samples with true label i and predicted label j.

---

Raises

• ValueError
  If input arrays are not the same shape, are empty, are not 1D, or contain NaN/Inf.

---

Example Usage

import numpy as np
from machinegnostics.metrics import confusion_matrix

# Example 1: Basic usage
y_true = [2, 0, 2, 2, 0, 1]
y_pred = [0, 0, 2, 2, 0, 2]
cm = confusion_matrix(y_true, y_pred)
print(cm)
# Output:
# array([[2, 0, 0],
#        [0, 0, 1],
#        [1, 0, 2]])

# Example 2: With custom labels
cm_custom = confusion_matrix(y_true, y_pred, labels=[0, 1, 2])
print(cm_custom)

---

Notes

• Compatibility: Accepts NumPy arrays, lists, or Pandas Series.
• Both y_true and y_pred must be 1D, have the same shape, and must not be empty or contain NaN/Inf.
• If labels is not provided, all unique labels in y_true and y_pred are used in sorted order.
• The confusion matrix is essential for computing other metrics such as accuracy, precision, recall, and F1-score.

---

Author: Nirmal Parmar
Date: 2025-09-24

---