贝叶斯网络在疾病预测诊断中的应用与优化

摘要

现今，国内外疫情形势依然严峻，新冠疫情已经严重影响人们正常的生活，甚至造成了全球经济的衰退化，疫情对经济产生的直接影响，是我们很直观可以看到、可以感受到的。比如说旅游景点关闭、餐厅不能正常营业、电影院关门等等。在全民接种疫苗的时代，新冠肺炎病毒也在不断的变异，对人民的生命和财产安全造成了很大的威胁。贝叶斯网络作为问题不确定推理的一个有效手段，它结合图论相关知识和概率论不仅可以完成问题的后验概率推理，还能很清楚地描绘出变量之间存在的依赖关系，具有很强的应用价值。因此，可以高效的帮助医护人员筛选较大概率患有新冠肺炎的患者。
    本文先以类似新冠肺炎症状患者为研究对象，提出基于贝叶斯网络的新冠肺炎检测方法 。基于目标人群的国家、年龄、性别、症状、等数据样本以及贝叶斯网络的参数学习，由此建立基于贝叶斯网络的疾病预测与优化预测模型 ，通过构建贝叶斯网络对人群患新冠肺炎的概率大小进行预测，以提高管理效率，帮助医护人员判断那些患者需要重点关注，降低新冠肺炎传播带来的危害。

关键词：医疗诊断、贝叶斯网络、新冠肺炎疫情

 

ABSTRACT

Today, the epidemic situation is still grim at home and abroad. COVID-19 has seriously affected people's normal life, and even caused the deterioration of the global economy. The direct impact of the epidemic on the economy is very intuitive to see and feel. For example, tourist attractions are closed, restaurants are not open normally, cinemas are closed, etc. The novel coronavirus pneumonia is also changing constantly in the era of universal vaccination, which poses a great threat to the safety of people's lives and property. As an effective means of problem uncertainty reasoning, Bayesian network combines the relevant knowledge of probability theory and graph theory. It can not only complete the posterior probability reasoning of the problem, but also clearly describe the dependency between variables, which has strong application value. Therefore, novel coronavirus pneumonia patients can be screened efficiently by health care workers.
Novel coronavirus pneumonia patients with novel coronavirus pneumonia were selected as the subjects. Novel coronavirus pneumonia novel coronavirus pneumonia prediction model based on Bayesian network is established based on gender, age, gender, symptom, data and Bayesian network parameters learning. The Bayesian network is used to predict the probability of new crowns pneumonia, so as to improve the management efficiency and help the medical staff to judge the patients' need to focus on the disease and reduce the harm caused by the spread of the new crown pneumonia.

Key words:medical diagnosis; Bayesian netw orks; COVID-19

一、绪论

（一）研究背景

新冠病毒是一种传染性极强的病毒，主要的传播途径是呼吸道飞沫传播和接触传播，因其传染性极强，当下全球各地都受到了来自新冠病毒不同程度上的影响。新冠病毒病发有着较强的潜伏期，感染新冠后，普通人很难第一时间察觉，从而会引发更多的传染。当前，如何有效诊断新冠肺炎成了亟待解决的问题。

（二）目的和意义

传统的诊断方法是核酸检测，但是由于新