Machine Learning Framework for Modeling and Predicting of Cyber Hacking Breaches

Abstract

Breaking down cyber incident data sets is a significant strategy for extending our comprehension of the advancement of the threat situation. This is a moderately new research subject, and numerous examinations stay to be finished. Right now, report a measurable investigation of a breach incident data set relating to 12 years of cyber hacking exercises that incorporate malware attacks. We show that, as opposed to the discoveries detailed in the writing, both hacking breach incident inter-arrival times and breach sizes ought to be displayed by stochastic procedures, instead of by disseminations since they show autocorrelations. At that point, we propose specific stochastic procedure models to, separately, fit the interarrival times and the breach sizes. We additionally show that these models can foresee the inter-arrival times and the breach sizes. So as to get further bits of knowledge into the development of hacking breach incidents, we direct both subjective and quantitative pattern examinations on the data set. We draw a lot of cyber security bits of knowledge, including that the threat of cyber hacks is to be sure deteriorating as far as their frequency, yet not regarding the extent of their harm.