A distribution model for data leakage prevention

In this paper we present a file distribution model aiming at data leakage prevention. Taking into consideration the guilt probability which is the probability of the user having leaked files, this model can select a file allocation plan with the least overlap between obtained file sets of users, as the consequence the model can find out leakage sources with high probability. So that measures can be taken for information security further. The simulation experiments reveal that the model can distinguish malicious users and detect the leak source effectively.