论文部分内容阅读
目的探索反向传播(BP)人工神经网络在国境卫生检疫领域中的应用研究。方法采用18×5×1结构的3层BP神经网络模型,对2007年到达宁波港国际航行船舶中的媒介阳性船舶170艘和对照船舶680艘进行数据训练和验证,并以建立的神经网络模型预测新到港的船舶外来媒介携带率。结果经过100次的迭代运算,训练过程的误判率为0.1647,验证过程的误判率为0.1824;训练过程的平均误差为0.3668,而验证过程的平均误差为0.4550。通过该神经网络模型预测船舶携带外来媒介情况与实际结果的符合率达到83.3%,预测效果良好。结论针对高度不确定的非线性系统,应用BP人工神经网络可实现相对精确的预测功能,为国境卫生检疫风险评估及预警方面的研究提供理论基础。
Objective To explore the application of back propagation (BP) artificial neural network in the field of frontier health and quarantine. Methods A 3-layer BP neural network model with 18 × 5 × 1 structure was used to train and verify the data of 170 medium-positive ships and 680 control ships in international voyages that arrived in Ningbo Port in 2007. Based on the established neural network model Predict portability of new arrivals to foreign carriers. Results After 100 iterations, the false positive rate of training was 0.1647, the false positive rate of validation was 0.1824, the average error of training was 0.3668, and the average error of verification was 0.4550. According to the neural network model, the coincidence rate of ship carrying foreign agent with the actual result reached 83.3%, and the prediction effect is good. Conclusion For highly uncertain nonlinear systems, the application of BP artificial neural network can achieve relatively accurate prediction function and provide a theoretical basis for the study of risk assessment and early warning of frontier health quarantine.