Commit 2024-05-20 22:00 abff007f
View on Github →feat: Embed a space into probability measures on it. (#11815)
Assuming a topological space X
is completely regular and T0, it can be embedded into ProbabilityMeasure X
. This shows in particular that for the Lévy-Prokhorov metrizability #11549 of ProbabilityMeasure X
, certain assumptions cannot be relaxed. For example we have:
import Mathlib.MeasureTheory.Measure.DiracProba
variable {Ω : Type*} [MeasurableSpace Ω] [TopologicalSpace Ω] [OpensMeasurableSpace Ω] [T0Space Ω] [CompletelyRegularSpace Ω]
open TopologicalSpace
example :
¬ MetrizableSpace Ω → ¬ MetrizableSpace (ProbabilityMeasure Ω) := by
intro badSpace goodProba
exact badSpace embedding_diracProba.metrizableSpace
example :
¬ PseudoMetrizableSpace Ω → ¬ PseudoMetrizableSpace (ProbabilityMeasure Ω) := by
intro badSpace goodProba
exact badSpace embedding_diracProba.pseudoMetrizableSpace
example :
¬ FirstCountableTopology Ω → ¬ FirstCountableTopology (ProbabilityMeasure Ω) := by
intro badSpace goodProba
exact badSpace embedding_diracProba.firstCountableTopology