Network Models for Data Science Theory, Algorithms, and Applications

There have been several attempts at incorporating real-world components into network generation and growth. Most attention has centered on trying to create a desired structure for the node-degree distribution, such as clustering and the power-law property. This chapter discusses the advantages and disadvantages of the “configuration” and the “expected-degree” models, and describes how the growth of a network can be formulated through the “preferential-attachment” and “random-copying” (or “duplication”) models.