多源信息下基于贝叶斯网络的综合性医院门诊量不确定性分析与预测

The outpatient visits in general hospitals are affected by many factors, which result in the randomness of the outpatients visits and disadvantageous to the use of medical resources.The purpose of this study is to set up bayesian models to forecast and evaluate the outpatient visits through multidisciplinary combination, and to measure the inherent rules between the factors and the outpatient visits qualitatively and quantitatively. At first, the factors influencing outpatient visits will be collected by many ways, and their physical meanings would be identified by the methods combining with physical and mathematical mining. For the uncertainty of the statistical distribution of factors, an adaptive probabilistic collocation based Kalman filter method will be put forward to simulate inversely the factor-parameters of outpatient visits. Secondly, An improved bayesian network combining fuzzy mathematics would be proposed to deduce the uncertainty of factors of outpatient visits and outpatients,as same as their inherent rules. The outcomes of the research, both the theories and methodologies definitely contribute to allocate resource reasonable in the general hospital and provide a warning signal of whether the population will has a larger scale disease. By the research, the ability of the data processing by bayesian network and the method of data inversion simulation would be expansion and enrichment.

综合性医院门诊量受多方面因素影响，容易造成门诊病人的随机性，不利于患者就诊和医院资源的合理利用。课题旨在融合多学科知识，通过贝叶斯网络建立门诊量预测与评估模型，定性与定量衡量门诊量影响因素之间、影响因素与门诊量之间的内在规律。首先多途径采集门诊量影响因素数据，再利用物理成因与数学挖掘相结合的手段识别影响因素的多层次物理意义；针对影响因素统计信息出现的不确定性，提出自适应概率配点的卡尔曼滤波法对参数进行反演模拟；在上述基础上，将贝叶斯网络与模糊数学相结合，提出改进贝叶斯门诊量预测模型对影响因素数据多样性问题进行处理，从而推断门诊量及其影响因素的不确定性和相互关系。项目的开展，将探讨贝叶斯预测模型处理多类型数据、不确定性参数反演模拟的方法，扩展和丰富医院门诊量监测和预警人群发生较大规模病情的理论框架体系和分析方法。

综合性医院门诊量受多方面因素影响，容易造成门诊病人的随机性，不利于患者就诊和医院资源的合理利用。项目采集了广西医科大学第一、第二附属医院、广西壮族自治区人民医院，南宁市第一人民医院门诊患者就医影响因素和门诊量数据，对数据采用智能算法手段作预处理与深度挖掘，探讨其多层次物理意义并针对门诊量影响因素提出了建议。依据门诊量数据特征，建立了适合门诊量数据特征的两个贝叶斯评估预测模型，为门诊量的预测提供了理论参考。

内容获取失败，请点击重试