Measurement of network-based and random meetings in social networks

Abstract

Social networks are created by the underlying behavior of the actors involved in them. Each actor hasinteractions with other actors in the network and these interactions decide whether a social relationship should developbetween them. Such interactions may occur due to meeting processes such as chance-based meetings or network-based(choice) meetings. Depending upon which of these two types of interactions plays a greater role in creation of links, asocial network shall evolve accordingly. This evolution shall result in the social network obtaining a suitable structureand certain unique features. The aim of this work is to determine the relative ratio of the meeting processes that existbetween different actors in a social network and their importance in understanding the procedure of network formation.This is achieved by selecting a suitable network genesis model. For this purpose, different models for network genesisare discussed in detail and their differences are highlighted through experimental results. Network genesis models arecompared and contrasted with other approaches available in the literature, such as simulation-based models and blockmodels. Performance measures to compare the results of the network genesis models with baselines are statistics ofnetworks recreated using the models. The socially generated networks studied here belong to various domains like ecommerce, electoral processes, social networking websites, peer to peer file-sharing websites, and Internet graphs. Theinsights obtained after analyzing these datasets by network genesis models are used for prescribing measures that couldensure continuous growth of these social networks and improve the benefits for the actors involved in them.