Step by Step
TP
True Positive — predicted positive, actually positive
The model predicted the positive class, and the true answer was actually positive — a correct prediction.
Example: a spam filter correctly flags an actual spam email as spam.
FP
False Positive — predicted positive, actually negative (Type I error)
The model predicted the positive class, but the true answer was actually negative — a "false alarm," also called a Type I error.
Example: a spam filter incorrectly flags a legitimate email as spam.
FN
False Negative — predicted negative, actually positive (Type II error)
The model predicted the negative class, but the true answer was actually positive — a "missed detection," also called a Type II error.
Example: a spam filter fails to catch an actual spam email, letting it through to the inbox.
TN
True Negative — predicted negative, actually negative
The model predicted the negative class, and the true answer was actually negative — another correct prediction.
Example: a spam filter correctly leaves a legitimate email in the inbox, not flagging it as spam.
Applied Walkthrough
1
A medical test predicts "disease" for a patient who does not actually have the disease.
2
Ask: which of the four boxes does this fall into? The model predicted positive, but the true answer was negative — a False Positive.
3
Contrast: if the same test predicted "no disease" for a patient who actually DOES have the disease, that would instead be a False Negative — a missed detection.
4
Every other classification metric (accuracy, precision, recall, F1, AUC) is calculated directly from these four confusion matrix numbers, which is why correctly identifying each box is the essential first step.
Exam Application
Exams test whether you can correctly classify a described prediction outcome into one of the four confusion matrix boxes, and whether you know the Type I/Type II error naming convention (False Positive = Type I, False Negative = Type II) that's commonly referenced in statistics-adjacent contexts.
⚠ Common Trap
The most common trap is confusing False Positive with False Negative. A useful check: the second word in each term ("Positive" or "Negative") tells you what the MODEL predicted, and whether it's "True" or "False" tells you whether that prediction was correct or not.
✓ Quick Self-Check
1. What does True Positive mean?
The model predicted positive, and the true answer was actually positive — a correct prediction.
Tap to reveal / hide
2. What does False Positive mean, and what is its alternate name?
The model predicted positive, but the true answer was actually negative — also called a Type I error, or false alarm.
Tap to reveal / hide
3. What does False Negative mean, and what is its alternate name?
The model predicted negative, but the true answer was actually positive — also called a Type II error, or missed detection.
Tap to reveal / hide
4. What does True Negative mean?
The model predicted negative, and the true answer was actually negative — a correct prediction.
Tap to reveal / hide
5. What do all other classification metrics (accuracy, precision, recall, F1, AUC) derive from?
The four confusion matrix numbers: TP, FP, FN, and TN.
Tap to reveal / hide