On the Randić index and its variants of network data

Abstract: Summary statistics play an important role in network data analysis. They can provide us with meaningful insight into the structure of a network. The Randi\'{c} index is one of the most popular network statistics that has been widely used for quantifying information of biological networks, chemical networks, pharmacologic networks, etc. A topic of current interest is to find bounds or limits of the Randi\'{c} index and its variants. A number of bounds of the indices are available in literature. Recently, there are several attempts to study the limits of the indices in the Erd\H{o}s-R\'{e}nyi random graph by simulation. In this paper, we shall derive the limits of the Randi\'{c} index and its variants of an inhomogeneous Erd\H{o}s-R\'{e}nyi random graph. Our results charaterize how network heterogeneity affects the indices and provide new insights about the Randi\'{c} index and its variants. Finally we apply the indices to several real-world networks.