项目名称: 基于大数据的基本养老保险多主体行为演化机理与宏观效应研究
项目编号: No.91546110
项目类型: 重大研究计划
立项/批准年度: 2016
项目学科: 管理科学
项目作者: 吕天阳
作者单位: 审计署计算机技术中心
项目金额: 43万元
中文摘要: 现代国家治理受到个人、企业、政府等多类海量主体行为的影响。在实践中,主体的微观行为常常聚合出预料之外又具潜在风险的宏观效应。大数据的累积,使得从微观角度剖析大量主体的行为互动和演化成为可能。基本养老保险涉及多类主体,其行为过程复杂,是具代表性的重大制度安排,将其作为切入点开展研究。.提出整合政务和互联网大数据,分析多类海量基本养老保险参保主体行为的真实演化过程,探索多因素交叠下主体行为的演化机理。具体包括:聚合社保、工商、税务等政务数据和互联网数据,形成数据分析基础;构建涉及百万参保主体的基于多属性节点的异质社会网络,作为互动性行为研究的基础;围绕“违规-监管”关系,依据对跨领域信息的交叉比对,识别参保主体的违规行为及其演化历程,剖析其影响因素;以演化博弈模型为基础,构建体现个人禀赋、互动关系、舆论环境等多重影响因素的演化模型;分析逃避参保、偷逃缴费、提前退休等行为的蔓延过程及其体制危害。
中文关键词: 大数据;基本养老保险;行为演化;政策仿真
英文摘要: Puts forward integrating government big data and internet big data, discovering the real evolution process of varied and large real pension participants, and exploring evolution mechanism of micro participants influenced by various overlapping factors. Specific contents are as follow: integrating internet big date and government big data from social security, commerce, telecommunication to offer big data base; constructs heterogeneous social network with multi-attribute node based on millions of pension participants, the purpose of which is to assist researches on interactions participants; chooses ‘misdeeds-supervision’ as the main subject, through comparing and verifying micro data from different fields, recognizes misdeeds and their evolution processes of participants, and analyses the influencing factors; based on evolutionary game models, constructs the evolutionary model containing overlapping influence by personal characteristics, interaction relationships, and public opinions; analyses the spreading process and the systematic crisis of avoiding participating, avoiding contributions, early retirement, etc.
英文关键词: big data;basic pension system ;action evolution;policy simulation