项目名称: 基于数据挖掘的地方政府绩效评估指标设计研究
项目编号: No.70873100
项目类型: 面上项目
立项/批准年度: 2009
项目学科: 矿业工程
项目作者: 卓越
作者单位: 厦门大学
项目金额: 20万元
中文摘要: 科学的绩效指标设计是政府绩效评估有效实施的核心环节,也是政府绩效评估理论和实践中的难点所在。导入数据挖掘技术,为政府绩效评估指标设计提供了一种全新的、有效的方法和技术。课题研究通过需求回应与效用机理分析,完成课题研究背景、研究概念、研究工具和研究对象等基础设施性铺垫工作;课题研究利用数据挖掘中的关联规则原理和集合理论中的加和性原理、数学中的向量原理来解决在海量数据中挖掘政府绩效评估指标时的可加性问题,疏通了运用数据挖掘技术导入政府绩效评估指标体系设计的关键技术节点;课题研究运用数据挖掘技术,历经数据清洗、数据仓库构建、指标设计、技术指标处理等一系列环节,选择典型研究对象,建构出一套较为完整的通用指标体系,完整展示三级指标内涵、计算方式、评分标准、评估数据来源等指标技术说明;课题研究运用主成分因子回归分析、决策树分类模型、聚类分析等不同的数据挖掘方法技术,选择典型研究对象,历经理论知识和业务理解、数据预处理、模型选择、结果分析与知识应用等一系列环节和步骤,探索性地为政府有关项目设计出行之有效的绩效评估指标,或为绩效评估指标设计提供有益的视角和思路。
中文关键词: 数据挖掘;政府绩效评估;指标设计;数据仓库
英文摘要: The scientific designation of performance indicator is one of the core and critical process of the implementation of government performance, also the difficulty that has to be conquered both in theory and practice of government performance evaluation. Data Mining technique provides a fire-new effective method for government performance evaluation. This study using demand response and utility mechanism analysis method to finish the foundational work of background analysis, research conception, instrument and objection. This research utilizing association rule of data mining technique, additive property of concentration theory and vector theory of mathematics to solve the additivity property problem when dating indicators of performance evaluation in tremendous figures and dredging the key technology node of applying data mining technique in the designation of index system of government performance evaluation. This discuss choosing representative objection of research to constructing a set of fairly integrated generally used indicator system, completely showing indicator technique explanation of connotation, computational formula, scale of marks and data resource by processes of data cleaning, database construction, indicator designation and technical indicator disposal. This study exploringly design effective performance evaluation indicators for related government projects by choosing representative research objection and processes of theoretical research, data pretreatment, model selecting, result analysis and knowledge application which would provide advantageous view and mentality of indicator designation for performance evaluation.
英文关键词: Data Mining;Government Performance Evaluation;Indicator Designation ;Database