A Comparative Study of Six Software Packages for Complex Network Research

Abstract

UCINET, Pajek, Networkx, iGraph, JUNG and statnet, are commonly used to perform analysis with complex network model. The scalability, and function coverage of these six software packages are assessed and compared. Some randomly generated datasets are used to evaluate the performance of these software packages with regard to input/output (I/O), basic graph algorithms, statistical metrics computation, graph generation, community detection, and visualization. A metric regarding both numbers of the nodes and the edges of complex networks, which is called Maximum Expected Network Processing Ability (MENPA), is proposed to measure the scalability of software packages. Empirical results show that these six software packages are complementary rather than competitive and the difference on the scalability among these six software packages may be attributed to the varieties in both of the programming languages and the network representations.