The continuous mapping theorem
ProvedFamousTheorems.continuous_mapping_theoremThe continuous mapping theorem. If random variables converge in distribution to and is continuous, then converges in distribution to .
It is one of the most used tools of asymptotic statistics. Combined with the central limit theorem and Slutsky's theorem it gives the limiting distributions of test statistics, such as limits of squared normal statistics.
Formalization note. Mathlib's MeasureTheory.TendstoInDistribution.continuous_comp. The variables may live on different probability spaces Ω i, convergence is along an arbitrary filter l, and TendstoInDistribution X l Z μ μ' means the laws converge weakly as probability measures on E.
Preamble
import Mathlib
Formal statement
namespace FamousTheorems
theorem continuous_mapping_theorem {ι E Ω' F : Type*} {Ω : ι → Type*} {m : ∀ i, MeasurableSpace (Ω i)} {μ : ∀ i, MeasureTheory.Measure (Ω i)}
[∀ i, MeasureTheory.IsProbabilityMeasure (μ i)] {m' : MeasurableSpace Ω'} {μ' : MeasureTheory.Measure Ω'}
[MeasureTheory.IsProbabilityMeasure μ'] {mE : MeasurableSpace E} {X : ∀ i, Ω i → E} {Z : Ω' → E}
{l : Filter ι} [TopologicalSpace E] [OpensMeasurableSpace E] [TopologicalSpace F] [MeasurableSpace F]
[BorelSpace F] {g : E → F} (hg : Continuous g) (h : MeasureTheory.TendstoInDistribution X l Z μ μ') :
MeasureTheory.TendstoInDistribution (fun n => g ∘ X n) l (g ∘ Z) μ μ' := by sorry
end FamousTheoremsSource
Listed in Mathlib's curated theorem manifests (docs/1000.yaml); formalized in Mathlib as `MeasureTheory.TendstoInDistribution.continuous_comp`. Proof here reduces to that Mathlib result.