Specificity, the true negative rate, selectivity

54 min
Block 15 — The classification metrics
Objective
define specificity as the fraction of actual negatives a rule leaves alone; establish specificity = 1 − FPR and why the ROC axis is the complement rather than the quantity; derive its invariance to prevalence and its monotone non-decrease in the threshold, the exact mirror of recall, and tabulate both rates on one score vector; give the three correct ways to compute it in a library that has no specificity_score; measure the central warning, that under imbalance specificity is almost always high and almost always uninformative, with an information-free rule measured at 0.998; work the alarm-volume arithmetic n(1 − p)(1 − s) and a costed screening program in which the specificity floor fixes the operating point; compute Youden's J as a threshold rule; compute one-vs-rest specificity and its inflation by the number of classes; compute confidence intervals; and measure what an unverified negative class does to a specificity figure.
Estimated duration
85 minutes
Prerequisites
chapters 011, 029, 030, 034, 050, 051, 052, 053, 054 and 055
Associated quizzes
056.1-quiz-definition-and-the-negative-row.md to 056.10-quiz-reporting-specificity.md

Preview — the rest of the lesson is for enrolled readers.

Already enrolled with a code?

Your access is tied to your account, not to this link. Sign in with the same email you used in class: your course is waiting, no need to enter the code again.

Sign inNo account yet? Create one
This lesson is part of the “Classification Metrics” module

The first modules of the course are open to everyone. For the rest you have three options: buy this course once and for all, subscribe, or enter the code handed out in class.

Are you a student on this course?

The code is tied to your account: sign in or create an account and it will be applied automatically when you come back.

No account yet? Create one