Prediction interval vs confidence interval — the difference.

medium

Answer

$CI: uncertainty in a parameter (μ, β, p) — shrinks with n.$
$Y_{\mathrm{new}}\; = \;{\mu}\; + \;{\varepsilon}$
$Practical: CI answers 'what's the average?' \to tightens with n.$
$PI answers 'where will the next point land?' \to bounded by irreducible variance.$

Check yourself — multiple choice

Same
CI: parameter uncertainty (shrinks with n); PI: parameter + irreducible noise on a new point (bounded by σ, doesn't shrink to 0)
PI is smaller
Only for regression

CI: parameter uncertainty; PI: new-point uncertainty including σ (wider, floor at σ).

#confidence-interval#regression

215 interview questions in this topic.