“…However, modelling the number of infected and susceptible individuals as continuous variables gives only poor approximations in small populations. In such cases, the system can be modelled as a Markov process with discrete population and time, where the probabilities of transmission from ðs; iÞ to ðs À 1; i þ 1Þ and ðs; i À 1Þ are given by bsi DT and gi DT, respectively (Witten and Poulter, 2007). Recently, network approaches to disease (Eubank et al, 2004;Liljeros et al, 2001;May and Lloyd, 2001;Newman, 2003;Nowak and Bangham, 1996;Verdasca et al, 2004;Witten and Poulter, 2007) and information (Kincaid, 2000;Kossinets and Watts, 2006;Valente, 2005;Watts and Strogatz, 1998) transmission have proven to allow more detailed predictions as they release the assumption of fully-mixed populations.…”