The Setting
Independent hyperedge percolation permits several vertices to be joined by a single open hyperedge. An ordered list of roots distinguishes clusters according to the first root they contain. Weighting such clusters by a function of their vertex sets gives a vector indexed by the roots.
A labelled hypergraph on a vertex set , with labels , is an assignment of a set to each ; empty hyperedges are permitted.✦ Give each label an opening probability . The Bernoulli hyperedge measure is the product measure under which the states are independent and .✦
For a forbidden vertex set , vertices and have open hyperedge adjacency when an open hyperedge disjoint from contains them both.✦ Open connectivity in after deletion of hyperedges meeting is connectivity by a finite chain of these adjacencies, including a chain of length zero.✦ For a vertex and a source set in , the cluster is the set of vertices connected to in using open hyperedges disjoint from , and the cluster union is the union of these clusters over vertices in , as in (1). (1) Write for the cluster without deletion.
Cluster Columns and Resampling
For roots , the earlier roots at index form the set in (2). For a vertex and a set , the admissible cluster domain is the family of vertex sets containing and disjoint from . (2)
The first root connection event in is the event that the cluster of after deletion of hyperedges meeting avoids the earlier-root set and contains , as in (3). (3) The weighted cluster column for an observable is the unnormalized expectation in (4). (4)✦ The unnormalized cluster matrix has the constant-observable columns specified in (5). (5)✦ The cluster coefficients are the vector defined in (6). (6)✦
For a root , a forbidden set , and an observable , the cluster expectation after deletion is the expected value of on the cluster of after deletion of hyperedges meeting , as in (7). (7) For a root and a source set in , conditional cluster resampling is the operator on observables defined by (8): for a vertex set , the forbidden set in the inner expectation is the cluster union from after deletion of hyperedges meeting . (8) For a root , source set , vertex set , and real observable on vertex sets, the residual cluster expectation is the expected correction weight assigned to when the weighted root cluster with weight is compared with its conditional cluster resampling at relative to .
The Claim
Theorem 1. Let be finite. Suppose that are distinct, that for every hyperedge label, and that is nonnegative and inclusion-increasing on . The cluster coefficients satisfy (9). (9)✦ For any ordered root list, probabilities , and real observable , the column decomposition is (10). (10)✦ Let , let for every hyperedge label, and let be nonnegative on . There is a constant for which (11) holds. (11)✦
In the paper, the result reads: for a finite hypergraph with independent hyperedge probabilities strictly below one and any ordered list of distinct roots , not necessarily all vertices, with denoting the random cluster of , set and , let be the conditional law of given on , and put , , and ; for every nonnegative inclusion-increasing on , the vector belongs to the nonnegative cone generated by with , and in particular is coordinatewise nonnegative.
What Is Checked
The statements concern Bernoulli product measures on labelled hypergraphs with a finite vertex set, their cluster columns and resampling operators; the coefficient inequality and the constant limit require for every label, with distinct ordered roots and a nonnegative increasing observable on for the former, and and a nonnegative observable on for the latter.