Confusion Matrix

Introduction

• Table layout to visualize performance of a classification algorithm
• Foundation for many other metrics, e.g. Precision and Recall

Examples

• True Positive (TP):
  • Actual = Cat
  • Predicted = Cat
• False Positive (FP)—Type I error:
  • Actual = Dog
  • Predicted = Cat
• True Negative (TN):
  • Actual = Dog
  • Predicted = !Cat
• False Negative (FN)—Type II error:
  • Actual = Cat
  • Predicted = !Cat