A Bayesian Network Learning Method with Easy Reasoning
Zhenxu Wu, Zhanpeng Yu, Fangchengyang Hou, Qingqing Sun · 2021 3rd International Conference on Applied Machine Learning (ICAML) · 2021
In connection with the current Bayesian network to the process of application, the reasoning efficiency is a little low, which can not be applied to the actual production and life. In order to improve the reasoning efficiency of the Bayesian, in this paper, an easy reasoning Bayesian network learning method based on reasoning complexity of Bayesian network is proposed. By adding the influence factors of reasoning efficiency into the original scoring function of Bayesian network, the influence factors of single reasoning and multiple reasoning are added into the learning process of Bayesian network. Based on the traditional hill-climbing method to learn Bayesian network, the Bayesian network association container is constructed. Under certain constraints, the easy reasoning Bayesian network is found. Experimental results show that the Bayesian network learning results of the proposed method can effectively reduce the reasoning complexity of Bayesian network, so as to improve the reasoning efficiency of Bayesian network in practical application.