突发传染病的小波神经网络预测  

Wavelet neural network predication for epidemic outbreak

作　　者：李晓龙[1] 田薇[1] 李晓毅[1] 

机构地区：[1]沈阳师范大学数学与系统科学学院,沈阳110034

出　　处：《沈阳师范大学学报（自然科学版）》2015年第3期392-395,共4页Journal of Shenyang Normal University:Natural Science Edition

基　　金：国家自然科学基金资助项目(10471096)

摘　　要：利用小波神经网络对突发传染病的预测进行研究。给出密度函数的小波估计的计算公式,提供了小波神经网络结构设计的理论框架。用小波函数作为隐层节点激活函数,神经网络连接权的大小由小波函数的系数确定,取数据库中的监控数据为训练样本,对小波神经网络进行训练学习,得到优化的神经网络。给出小波神经网络学习过程和具体步骤,用小波神经网络对突发传染病历史数据库中的已知数据,进行未知密度函数的小波估计,得到相应的小波估计函数和分布函数,在显著性水平下做拟合检验,构造激活函数,得到输出结果,进而进行预测,验证其有效性和可行性,最后总结问题的关键和今后研究的方向。Using wavelet neural network to predict outbreaks of infectious diseases were studied.The calculation formula of the wavelet estimation of density function is given,which provides a theoretical framework for the structural design of the wavelet neural network.Using the wavelet function as the activation function of the hidden layer nodes,the connection weights of the neural network is determined by the coefficient of the wavelet function.The monitoring data in the database is a training sample,and the wavelet neural network is trained to learn and get optimized neural network.The learning process and the concrete steps of the wavelet neural network are presented.Using the wavelet neural network to the known data in the history database of the burst infectious disease,the wavelet estimation of the unknown density function is carried out,and the corresponding wavelet function and distribution function are obtained,Under the significance level,the fitting test is done,and the activation function is constructed,and the output results are obtained,and then the validity and feasibility of the research is verified.Finally,we summarize key issues and future directions of research.

关 键 词：小波神经网络 密度估计 拟合检验 分类预测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]