Probability of infectious nodes in backward time

Computer worms have infected millions of computers since 1980s. In this paper, we consider the problem of estimating the probability of infection of each node in time steps of worm propagation backwards in time with Bayesian networks. The probability of infection of a node at each time depends on the out_degree of the node and the number of infectious nodes at that time. We assume to have prior knowledge of worm infection parameters. We also assume to have a snapshot of the network in which the number of susceptible, infectious and removed nodes are known. We repeatedly collect the out_degree of nodes in current sliding window. It has already been shown that, the number of infectious nodes can be approximated probabilistically backwards in time, through three Back-to-Origin models, i.e. deterministic and stochastic Back-to-Origin models and stochastic Back-to-Origin Markov model. Our method estimates origin node and epidemic ripple probabilistically.