Day 49 in MIT Sloan Fellows Class 2023, Advanced Data Analytics and Machine Learning in Finance 2, Confusion matrix
The confusion matrix is super confusing.
Everytime I forget the definition of TP, FP, etc, and metrics as well.
- True positives (TP) are positive outcomes that the model predicted correctly.
- True Negatives (TN) are negative outcomes that the model predicted correctly.
- False Negatives (FN) are negative outcomes that the model predicted incorrectly.
- False Positives (FP) are positive outcomes that the model predicted incorrectly.
So if it is a cancer test, TP means we predict cancer correctly.
TN means we did not predict cancer and patient actually does not have cancer.
FP is we predict a patient has cancer, but actually he or she does not have.
FN is we predict a patient is healthy, but actually he or she has a cancer.
- Sensitivity: How many positive people you can predict from positive cohort?
- Accuracy: How many you correctly predicted in all the cohort?