基本养老保险基金失衡风险预警方法的选择研究——基于新疆数据的实证分析  

The Choice of Imbalances Risk Early Warning Method in Basic Pension Insurance Fund——Xinjiang-based Data Analysis

作　　者：杨一帆[1] 朱衡[1] 

机构地区：[1]西南交通大学公共管理学院,四川成都610031

出　　处：《新疆大学学报（自然科学版）》2013年第4期490-496,共7页Journal of Xinjiang University(Natural Science Edition)

基　　金：国家自然科学基金(71303195);教育部人文社科研究项目(11YJC630256)的支持

摘　　要：纵观现下中国基本养老保险的发展,如何建立可持续发展的养老金制度,实现养老金的长期财务平衡,降低养老金运行中的风险等问题突出的摆在社会面前.本文尝试通过BP神经网络构建的基本养老金运行风险预警系统预测结果与多元回归模型的比较,进行基本养老金失衡风险预警方法的选择研究.在建立基金运行的风险指标体系基础上,搜集整理源于新疆的具体数据,首先进行多元回归模型的基金风险预警结果分析,然后通过MATLAB实现BP神经网络模型的反复训练、学习,并通过预警值与实际值的比较,检验BP神经网络预警模型的有效性,最后通过两种模型结果的比较,选择出BP神经网络预警方法更有优势.Throughout the development of the basic pension which brings many lessons for China, the problems of how to build a sustainable pension system to achieve long-term financial balance of the pension, at the same time to reduce operational risks in pension face everybody. Through the comparison of forecast results of BP neural network and multiple regression model the article attempts to make the choice of risk early warning method of pension. We collected specific data from Xinjiang in the establishment of a pension＇s risk indicator system running, and then analyze results of the fund risk warning of multiple regression model. What is more, the article uses MATLAB to achieve BP neural network model of repeated training, learning, and then through the warning value and the actual value comparison to identify the validity of BP neural network prediction. Finally through the comparison of the results in two models to choose BP neural network as a better model for Warning.

关 键 词：基本养老金 多元回归 BP神经网络 

分 类 号：F840.612[经济管理—保险]