Probabilistic fire models are categorized into network, statistical, simulation, and behavioral models. Network models use graph theory to represent the fire spread and the fire protection system in a building. Statistical models use historical data and probability distributions to estimate the likelihood and consequences of fire scenarios. Simulation models use mathematical equations and numerical methods to describe the physical and chemical processes of fire and its effects on the environment and the occupants. Behavioral models use psychological and sociological theories to predict the human response and evacuation behavior in case of fire. Network models are not a common category of probabilistic fire models, and they are not mentioned in the sources provided by the user. Therefore, network models are the correct answer.

References:Probabilistic Fire Simulation Assessment Using Simplified Model and Zone Modelling of a Kitchen Fire Scenario;Computer Fire Models for Fire Investigation and Reconstruction;Fire Behavior