传染病的自组织神经网络预测研究

Study on prediction of Infectious Diseases Based on Self-Organizing Neural Network Method

针对传染病发病趋势的预测需要考虑许多因素的特点,提出了用改进的自组织神经网络方法对传染病发病趋势进行预测。改进的自组织神经网络的算法用免疫克隆选择算法的克隆算子和变异算子去改进自组织神经网络算法中的邻域大小和权值调整规则,使每个神经元的权值学习率和邻域大小随神经元的亲和力发生变化,从而保证网络在很大概率上收敛到全局最优,并克服了自组织神经网络分类效果受输入次序影响的不足。最后实例分析表明了该方法的实用性和有效性。

This paper presents an improved self-organizing neural network model to predict the trend of infectious diseases because a number of factors should be considered.The improved self-organizing neural network model use clonal operator and mutation operator of immune clonal select algorithm to improve the neighborhood size and weight adjustment rules of self-organizing neural network algorithm,which changes the weight learning rate of each neuron and the neighborhood size of the neurons with the affinity of neurons,so it is ensured that network convergence to the global optimum in a great probability,and it overcomes the shortage that the classification effect of self-organizing neural network is prone to affect by input sequence.Finally,an example of the model shows the usefulness and effectiveness.