A Bayesian Network CPT Construction Method Integrating Expert Inference Informatio

Bayesian network is an important method in the field of artificial intelligence,it has become a hot spot with a solid theoretical foundation,the natural expression,the flexible reasoning capacity and the decision-making mechanisms to facilitate research in recent years.Conditional probability table(CPT)is the key issue in the process of constructing Bayesian network model,which can be determined by machine learning methods to make a frequency statistics based on the experimental data or observation data,also can be determined the knowledge and experience of experts subjective estimates.The estimation method combined with expert priori knowledge can effectively fuse the knowledge experience and test data and also can ignore minor relationships and highlight the main contradiction to effectively prevent "over learning",which is considered an important tool for Bayesian network learning.This paper explored a new Bayesian network CPT construction method by introducing experts knowledge to Bayesian Networks CPT build process.First,concerning the problem of lacking completeness and accuracy in the individuals inference information and lacking scientific in the overall integration results,which existed in the process of inferring conditional probability tables in Bayesian network according to experts knowledge,this paper presented a method based on the DS/AHP theory for deriving optimal conditional probability from the inferred information.First the inferred information extraction mechanism was proposed to make judgment objects more intuitive and judgment modes more perfect by introducing the knowledge matrix of the DS/AHP method.Finally the traditional method and the presented method were applied to infer the missing conditional probability table in the same Bayesian network.It is shown in the numerical comparison analysis that the calculation efficiency and accuracy can be improved through the proposed method.Meanwhile,the presented method is illustrated to be scientific.Secondly,In order to solve the problem of lacking effectiveness in the individuals inference and scientific in the overall integration results,which existed in the conditional probability tables in Bayesian network according to experts knowledge inference,the inferred information reduction mechanism is proposed to make judgment objects more intuitive and judgment modes more simple.Then the pair-wised judgment matrix in the analytic hierarchy process(AHP)is used as an extraction means and information carriers of the subjective conditional probability.After that,the experts relative inference method for deriving optimal conditional probability from macro and micro information is constructed,and the construction process of Bayesian network CPT is proposed.Finally a data comparison analysis is employed to prove the present method to be scientific.Finally,the application of the two proposed Bayesian network CPT construction method in multimorbidity Bayesian network CPT modeling were researched.Building multimorbidity Bayesian network with experts knowledge was analyzed in the field of health care based on the proposed Bayesian network CPT construction method,multimorbidity Bayesian network CPT construction methods based on the DS/AHP principle and integration experts relatively inference were proposed,which provided a scientific reference to carry out the prediction and diagnosis based on multimorbidity Bayesian network in the health care field.The presented method was also illustrated to be applicable and feasible,then provided a new idea to build a Bayesian network CPT based on experts knowledge.