The Giry monad #
Let X be a measurable space. The collection of all measures on X again forms a measurable space. This construction forms a monad on measurable spaces and measurable functions, called the Giry monad.
Note that most sources use the term "Giry monad" for the restriction to probability measures. Here we include all measures on X.
See also MeasureTheory/Category/MeasCat.lean
, containing an upgrade of the type-level
monad to an honest monad of the functor measure : MeasCat ⥤ MeasCat
.
References #
Tags #
giry monad
Measurability structure on Measure
: Measures are measurable w.r.t. all projections
Equations
- MeasureTheory.Measure.instMeasurableSpace = ⨆ (s : Set α), ⨆ (_ : MeasurableSet s), MeasurableSpace.comap (fun (μ : MeasureTheory.Measure α) => μ s) (borel ENNReal)
Equations
- ⋯ = ⋯
Monadic join on Measure
in the category of measurable spaces and measurable
functions.
Equations
- m.join = MeasureTheory.Measure.ofMeasurable (fun (s : Set α) (x : MeasurableSet s) => ∫⁻ (μ : MeasureTheory.Measure α), μ s ∂m) ⋯ ⋯
Instances For
Monadic bind on Measure
, only works in the category of measurable spaces and measurable
functions. When the function f
is not measurable the result is not well defined.
Equations
- m.bind f = (MeasureTheory.Measure.map f m).join