DEPENDENCE OF COMPUTER VIRUS PREVALENCE ON NETWORK STRUCTURE — STOCHASTIC MODELING APPROACH