基于神经网络的城镇在岗职工平均工资预测  

Forecasting the Average Wage of Urban Workers Based on Neural Network

作　　者：梅端[1] 周会会[1] 

机构地区：[1]广东海洋大学理学院

出　　处：《河北工程技术高等专科学校学报》2016年第1期36-38,共3页Journal of Hebei Engineering and Technical College Quarterly

基　　金：广东省高校创新强校工程项目;项目编号2014KQNCX080

摘　　要：由于在岗职工的平均工资不仅能反映人民的生活水平,也是企业退休职工的养老金、社保基金收缴比例以及个税征收比例等主要指标的参考标准,因此对其进行预测就显得尤为重要。为了提高在岗职工平均工资的预测精度,文中分别采用了BP神经网络和广义回归神经网络(GRNN)的方法,利用Matlab2010a对城镇在岗职工的平均工资进行模拟和预测。其中,用BP神经网络预测出的平均工资与实际值十分接近,相对误差能控制到2%以内,因此,利用BP神经网络对平均工资进行预测是可行的。The average wage of workers can not only reflect the people＇s living standard,but also can be used as a reference standard for enterprise retirees pension,social security funds,tax collection ratio of tax rates and other key indicators.So it is particularly important to predict the average wage of workers.In order to improve the prediction accuracy of the average wage of workers,this paper uses two methods which are BP neural network and generalized regression neural networks（GRNN）to simulate and predict the average wage of urban workers,on the basis of Matlab2010 a,respectively.Among them,the predicted result of BP neural network is close to the true value,and the relative error can be controlled within 2%.Furthermore,the method of BP neural network has the higher accuracy.Therefore,taking the BP neural network as a tool to predict the average wage is feasible.

关 键 词：BP神经网络 广义回归神经网络 平均工资 预测 

分 类 号：F244[经济管理—劳动经济]