Exchangeable graph laws correspond to mixing measures on graphon space #
The mixture map from probability measures on the graphon space over the unit interval to
exchangeable graph laws is injective (mixtureExchangeableLaw_injective): a mixture law determines
its mixing measure. Together with existence (exists_mixtureExchangeableLaw_eq) this makes the
mixture map a bijection, packaged as mixtureExchangeableLawEquiv. This is the Diaconis–Janson
correspondence at the level of the finite marginals. Uniqueness holds on the graphon quotient
GraphonSpaceI, not among graphon representatives.
Main definitions #
TauCeti.DenseGraphLimits.mixtureExchangeableLawEquiv— the mixture map, as an equivalence between probability measures onGraphonSpaceIand exchangeable graph laws.
Main results #
TauCeti.DenseGraphLimits.mixtureExchangeableLaw_injective— a mixture law determines its mixing measure.TauCeti.DenseGraphLimits.mixtureExchangeableLawEquiv_apply— the equivalence is the mixture map.
References #
- P. Diaconis, S. Janson, Graph limits and exchangeable random graphs, Rend. Mat. Appl. (7) 28 (2008), 33–61, Section 5.
- L. Lovász, Large Networks and Graph Limits, AMS Colloquium Publications 60 (2012), Section 11.3.
Uniqueness of the mixing measure. Two probability measures on GraphonSpaceI with the same
mixture law are equal.
The Diaconis–Janson correspondence. The mixture map is an equivalence between probability
measures on GraphonSpaceI and exchangeable graph laws.
Equations
Instances For
The forward map of mixtureExchangeableLawEquiv is the mixture map.