地方政府绩效评估指标体系挖掘过程——以江苏四市为例示  

The Process to Mine the Local Government Performance Evaluation Indicators by DM:An Pilot Study in the Four Municipal Governments in Jiangsu Province

作　　者：尚虎平[1] 

机构地区：[1]兰州大学管理学院,兰州730000

出　　处：《甘肃行政学院学报》2012年第6期23-42,124,共20页Journal of Gansu Administration Institute

基　　金：国家社科基金重大项目(11&ZD057);国家教育部人文社会科学研究青年项目(编号:09YJC630126);国家自然科学基金(编号:70873100)的阶段性成果

摘　　要：在建成江苏四市政府绩效评估指标数据仓库之后,就有了数据挖掘的基础。指标体系的设计实际上是一个在诸多可能的备选项中分层赋权聚类的过程,本文正是基于此思路,利用Microsoft SQL Server 2005软件中的Microsoft Visual Studio所提供的"神经网络"和"聚类分析"两个数据挖掘工具(技术)展开分层赋权聚类并将其中因为假设值为1而造成的"过拟合"正则化,从而产生本研究的结果——一套基于海量数据挖掘的、探索性的地方政府绩效评估指标体系。本文呈现的是以江苏四市为例示如何对地方政府绩效评估指标体系的三级指标数据进行挖掘的过程。In such a tremendously large country like China,there are substantial differences on nature,geographic features,history backgrounds and social realities among the varied regions.Because of the heterogeneity,the designing of the local government performance evaluation indicators is a systematical engineering which could be affected by various factors.But we can solve the problem by data mining and establishing a government performance evaluation indictor data warehouse.In this paper,we use the 'standard proportion' to develop a new way to choose the study objects among the great many governments of varied level in China.Then we bring the data mining into the study of the local government performance evaluation indicators design and take advantage of the Microsoft SQL Server 2005 as the DM tool.We establish a government performance evaluation indictor data warehouse of the four municipal governments(Suzhou,Nanjing,Yancheng,Xuzhou) in Jiangsu Province by processing the data of National Statistical Yearbook, Local Statistical Yearbook,Sector Statistics Yearbook,and government departments'work plans & programs,etc.

关 键 词：地方政府绩效评估 数据挖掘 绩效评估指标体系 数据仓库 

分 类 号：D035[政治法律—政治学]