{"id":"arxiv/2607.26624","url":"https://arxiv.org/abs/2607.26624","tags":["math.ST"],"score":0,"title":"Hub Neighbor-Degree Diagnostics for Sparse Random Graphs","domain":"math.ST","source":"arxiv","authors":["Qian Hui","Tiandong Wang"],"reviews":0,"subject":"math.ST","summary":"Networks with nearly identical degree distributions can place their hubs in sharply different neighborhoods. We develop a model diagnostic based on the mean degree of the neighbors of a degree-$k$ vertex. Under rank-one inhomogeneous random graphs, this statistic has degree-invariant centering and $k^{-1/2}$ fluctuations. Under non-rank-one kernels, posterior uncertainty about the root type can instead determine both centering and scale. Under linear preferential attachment, the statistic grows as $(m+δ)\\log k$. We turn these model-specific limits into goodness-of-fit tests for specified…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-07-29","reactions":{}}