摘要
本文首先构建失业综合指数及其测算指标体系，用来客观地反映失业状况；其次通过文献研究、发放问卷调查等分析方法，并在考虑数据可获得性、可靠性等前提下，初步整理出影响失业的相关因素，以2000～2012年各季度对应的数据，对失业综合指数及其影响因素进行多元线性逐步回归分析，从中筛选出影响显著的因素，构建失业综合指数预测模型。同时基于线性相关分析筛选的结果，构建BP神经网络预测模型，同样对失业状况进行预测，并与多元回归预测模型的预测结论进行比较，结果发现后者预测性能高于前者。
Firstly , this paper structured the composite index of unemployment （CIU ） , which could reflect the unemployment situa-tion objectively ;secondly , through the way of literature review , questionnaire surveys , and taking into account the availability and relia-bility of the data , the writer sort out the relevant factors which affecting unemployment preliminarily . Finally , based on the data during the years of 2000~2012 , with SPSS17.0 software , multiple linear stepwise regression method have been applied to analysis the relationship between CIU and its influencing factors , and the key factors which affect CIU most significantly were extracted to construct the multiple re-gression forecast model of CIU . Meanwhile ,based on the results of the linear correlation analysis ,and through software of MATLAB R12a , the BP neural network prediction model were constructed too , and comparing with multiple regression prediction model , the research con-clusions showed that the latter prediction model performance better than the former .
出处
《工业技术经济》
北大核心
2014年第2期103-112,共10页
Journal of Industrial Technological Economics
基金
教育部人文社会科学研究青年基金项目(项目编号:11YJC880089)
浙江省教育厅科研项目(项目编号:Y201122004)
关键词
失业综合指数
多元线性回归
BP神经网络
预测模型
composite index of unemployment
multiple linear regression
BP neural network
forecast model
作者简介
沈国琪，湖州师范学院社会发展与管理学院讲师，南京航空航天大学经济与管理学院管理学博士。研究方向：人力资源开发与管理。
陈万明，南京航空航天大学经济与管理学院教授，博士生导师。研究方向：人力资源管理。
