Commit 2023-08-02 05:14 5cbb0ad6
View on Github →feat(Probability/Independence): define independence wrt a kernel and a measure (#6106)
We introduce a new notion of independence with respect to a kernel and a measure. The plan is to eventually express both independence and conditional independence as particular cases of this new notion (see #6098).
Two sigma-algebras m
and m'
are said to be independent with respect to a kernel κ
and a measure μ
if for all m
-measurable sets t₁
and m'
-measurable sets t₂
, ∀ᵐ a ∂μ, κ a (t₁ ∩ t₂) = κ a t₁ * κ a t₂
.
Independence is the special case where κ
is a constant kernel. Conditional independence can be defined by using the conditional expectation kernel condexpKernel
.