Kernel security level software verification and confirmation method and system based on Bayesian belief network

The invention discloses a kernel security level software verification and confirmation method and system based on a Bayesian belief network. The method comprises the following steps: constructing an evaluation network model of current kernel security level software; for each evaluation network model, obtaining a relative importance degree value of each evaluation person to every two nodes under each target node in the evaluation network model; weighting by using the weight corresponding to each evaluator to obtain a relative importance degree weighted value, and constructing a plurality of judgment matrixes by using the relative importance degree weighted value; performing consistency verification on each judgment matrix; if the judgment matrix passes the verification, obtaining a reliability evaluation result of each node of the bottom layer by each evaluator, and weighting the reliability evaluation result by using a weight corresponding to the evaluator to obtain a reliability probability of each node; cal