Quantification of data extraction noise in probabilistic Boolean Network modeling

Probabilistic Boolean Networks have served as the main model for studying the application of optimal intervention strategies to favorably affect system dynamics. The errors originating in the data extraction or network inference process prevent the accurate estimation of the state transition probabilities of the network. The mathematical characterization of the uncertainties will enable us to analyze the performance of intervention strategies derived without considering the uncertainties and assist in the design of control policies robust to those uncertainties. In this paper, we will quantify the errors due to data extraction noise and discretization and their effects on the state transition and steady state probabilities of the probabilistic Boolean network.