基于人工神经网络的非稳态噪声性耳聋发病率预测模型的研究  

Research on forecasting model of incidence rate of deafness induced by instable state noise based on BP neural network

作　　者：严雁翎 何小新[2] 

机构地区：[1]武汉市劳动卫生职业病防治院,湖北武汉430071 [2]武汉大学机械系,430072

出　　处：《中国职业医学》2003年第6期25-27,共3页China Occupational Medicine

摘　　要：目的　提高非稳态噪声性耳聋发病率预测精度。方法　根据噪声性耳聋的发病率与其影响因素之间存在的映射关系 ,建立反向传播神经网络 (BP神经网络 )模型 ,用于非稳态噪声性耳聋发病率的预测。结果　得到由 2个输入单元、1个输出单元、9个隐含单元组成的BP神经网络非稳态噪声性耳聋发病率预测模型 ,预测值最大相对误差和平均误差都低于传统预测模型 ,达到满意效果。结论　BP神经网络模型作为噪声性耳聋早期预测的工具 。Objective To improve the forecasting precision of the incidence rate of deafness induced by instable state noise. Methods According to the mapping relation between the incidence rate of the noise deafness and their factors, the model of BP neural network was established to forecast the incidence rate of deafness induced by instable state noise. Results The model of BP neural network for forecasting the incidence rate of deafness induced by instable state noise would be composed of two input units and one output unit and 9 implicit units. The relative error and the average error of the forecasting values appeared to be less than that of the traditional forecasting model. The effect is greatly satisfying. Conclusion The BP neural network model would be valuable to be popularized as an important tool to forecast deafness induced by instable state noise at its early stage.

关 键 词：非稳态噪声 噪声性耳聋 预测 BP神经网络 

分 类 号：R135.3[医药卫生—劳动卫生]