A General Framework for Evolving Network Analysis
Abstract
Many efforts have been given to model physical systems involving a large number of interacting constituents. Such systems are commonly called complex networks for their complex behaviors demonstrated at the system, global level. As a network evolves its constituents (or nodes) and associated links would either increase or decrease or both. It is a challenge to extract the specifics that underlie the evolution of a network or indicate the addition and/or removal of links in time. Many evolving network analysis algorithms are available. However, the majority of these algorithms assume network evolution can be fully comprehended using a snapshot of the network as it progresses. Without retaining the time information would inevitably obscure the dynamics of the evolving network and misinterpret its behaviors. A general framework viable for describing evolving networks is developed in the thesis. The framework incorporates the two-dimensional discrete wavelet transform (2DWT) with the tensor factorization method to extract features indicative of the development of a network in the time-frequency domain. The general framework model is evaluated against six benchmark algorithms using five different real-world evolving network datasets to demonstrate its feasibility in achieving a high level of link prediction.
Citation
Wu, Bin (2018). A General Framework for Evolving Network Analysis. Master's thesis, Texas A & M University. Available electronically from https : / /hdl .handle .net /1969 .1 /174585.