System evaluation and detection method by Bayesian model

Abstract

The invention discloses a system evaluation and detection method by a Bayesian model. The method includes: step A, establishing a logic relationship and an event relationship between a software system and a database therein after variable and relationship conversion; step B, establishing a fault tree model for a logic model according to the step A; step C, establishing a Bayesian network based on the fault tree model; step D, conducting system evaluation and detection based on the Bayesian network. The system evaluation and detection method by the Bayesian model has the advantages of low computational complexity, low computational difficulty, simplicity and practicality when used for system evaluation and detection by the Bayesian model.

Description

Technical field [0001] The present invention relates to the field of computer software theory and software system research, in particular to a method for system evaluation and detection using Bayesian models. Background technique [0002] The definition of Bayesian network, Bayesian network, also known as Bayesian belief network, is a directed graphic description of probability relations, and it provides a way to visualize the intuitive map of knowledge. A Bayesian network is a Directed Acyclic Graph (DAG), which consists of nodes representing variables and directed edges connecting these nodes. Among them, nodes represent variables in the universe of discourse, and directed arcs represent the relationship between variables (that is, influence probability). The uncertainty knowledge is expressed through graphs, and the dependence of local conditions can be expressed in the model through the annotation of conditional probability distribution (CPD). Sex. [0003] It can be seen fro...

Images