摘要
[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- regularization BP neural network model of bacillary dysentery for the disease control. [Methods] The forecasting model for bacillary dysentery was established by using neural network toolbox of Matlab 7.2 package. The meteorological factors,data of social economic and incidence data of bacillary dysentery in Fuzhou from 1987 to 2006 were analyzed. The effectiveness of established forecasting model was tested by using data in 2007. [Results] After neural network training, the error of performance decreased and the correlation coefficient was 0. 842. The efficiency of forecasting for bacillary dysentery was 91.7 %. [Conclusion] Bayesianregularization BP neural network model is feasible to analyze the relationship of meteorological factors and incidence of bacillary dysentery. Bayesian-regularization BP neural network model may be used as a new effective method for forecasting incidence of bacillary dysentery.
出处
《海峡预防医学杂志》
CAS
2009年第2期9-11,共3页
Strait Journal of Preventive Medicine
关键词
细菌性痢疾
气象要素
贝叶斯正规化
BP神经网络
B acillary Dysentery
Meteorological Factors
Bayesian- regularization
BP Neural Network
