论文部分内容阅读
[目的]探索贝叶斯正规化BP神经网络在细菌性痢疾预测模型中的应用,为菌痢的预防控制措施提供科学依据。[方法]用Matlab 7.2软件包中的神经网络工具箱,以福州市城区1987—2006年的气象要素、社会经济资料和菌痢发病率数据进行分析,建立福州市城区菌痢流行的贝叶斯正规化BP神经网络模型,并以2007年的资料验证其预测效果。[结果]神经网络经学习和训练,训练误差下降并趋于稳定,回代相关系数为0.842,预测成功率为91.7%。[结论]贝叶斯正规化BP神经网络在气象要素与菌痢发病率之间建模是可行的,能作为预测菌痢流行的一种新方法。
[Objective] To explore the application of Bayesian regularized BP neural network in the prediction model of bacillary dysentery and provide a scientific basis for the prevention and control measures of bacillary dysentery. [Methods] The neural network toolbox in Matlab 7.2 software package was used to analyze the meteorological factors, socio-economic data and the incidence of dysentery in Fuzhou City from 1987 to 2006 to establish the epidemic Bayesian The normalized BP neural network model is normalized, and the data of 2007 are used to verify its prediction effect. [Result] After learning and training neural network, the training error decreased and stabilized. The correlation coefficient was 0.842 and the prediction success rate was 91.7%. [Conclusion] The Bayesian normalized BP neural network is feasible to model the relationship between the meteorological factors and the incidence of dysentery and could be used as a new method to predict the prevalence of bacillary dysentery.