基于聚类的数据挖掘技术在税源专业化管理中的应用

发布时间：2018-03-22 09:43

本文选题：聚类　切入点：税源专业化管理　出处：《集美大学》2017年硕士论文　论文类型：学位论文

【摘要】：继云计算后,大数据(Big Data)成为信息技术领域最为热门的概念之一。根据中国互联网数据中心的报告,2012年全球的数据总量2.7ZB(1ZB=106PB,1PB=106GB),而加上互联网和移动互联网,各类记载了个人购物、位置等方面的相关数据更是海量。所以如何利用这些数据,成为了日益突出的问题。作为为国聚财、为民收税的税务部门,同样存在着涉税数据缺乏深度挖掘利用的问题。随着“金税三期”的上线,税务部门的涉税数据从省级集中变成了全国数据的大集中,税务部门的数据库中存在海量的外部和内部数据。但是税务部门对这些海量数据的分析与处理局限于汇总、查询、对比等基础性的应用,无法对数据的深层信息进行挖掘。本文力图在税务部门的工作中引进数据挖掘技术,从而让税务部门的决策更加具有科学性并提高纳税服务质量,同时将研究的重心放到了聚类分析在税源专业化管理的应用上。税源专业化管理是基础的税务管理工作,掌控全面的税源信息,实行严谨的专业化管理方法,能够充分保证税务部门税收的应收尽收。聚类分析是数据挖掘技术的一种,目前已经在商业、医疗等领域广泛应用。将聚类分析用于税源专业化管理中,能够充分地提升税源分类的有效性,同时可以使税务人员能够科学地制定税源管理策略,提高税源专业化管理的效率,并有助于推动“放管服”改革不断深入。本文在总结了国内和国外研究与实践的基础上,梳理了税源专业化管理、大数据、数据挖掘技术和聚类算法等方面的相关知识,然后把各方面的知识综合起来,提出聚类算法在税源专业化管理应用的存在问题与对策研究,并研究设计了聚类算法在税源专业化管理应用的模型。继而用SPSS Clementine20.0软件,同时依照数据挖掘的流程建立了标准模型,通过文章选用的指标体系使用K-means聚类方法对WIND分类下的日常消费行业196家公司进行了分析,论证了聚类算法在税源专业化管理应用中的适用性。
[Abstract]:Following cloud computing, big data has become one of the hottest concepts in the information technology world. According to the China Internet data Center, the global data volume in 2012 was 2.7ZBP1ZBP106PBP106GB, while the Internet and the mobile Internet recorded individual shopping. Location and other related data are even more massive. So how to use these data has become an increasingly prominent problem. As a tax department collecting money for the country and collecting taxes for the people, There is also the problem of the lack of deep mining and utilization of tax-related data. With the launch of the third issue of the Gold tax, tax data of the tax authorities have changed from a provincial concentration to a large concentration of national data. There are huge amounts of external and internal data in the database of tax departments. However, the analysis and processing of these huge amounts of data by tax authorities are limited to basic applications such as summary, query, contrast, etc. This paper tries to introduce the data mining technology into the work of the tax department so as to make the decision of the tax department more scientific and improve the quality of tax service. At the same time, the focus of the research is on the application of cluster analysis in the specialized management of tax sources. The specialized management of tax sources is the basic work of tax administration, which controls the comprehensive information of tax sources and implements rigorous specialized management methods. Cluster analysis is a kind of data mining technology, which has been widely used in commercial, medical and other fields. Cluster analysis is used in specialized management of tax sources. It can fully improve the effectiveness of tax source classification, at the same time, can enable tax personnel to scientifically formulate tax source management strategies, improve the efficiency of specialized management of tax sources, On the basis of summarizing the research and practice at home and abroad, this paper combs the relevant knowledge of specialized management of tax sources, big data, data mining technology and clustering algorithm, etc. Then, by synthesizing all aspects of knowledge, the paper puts forward the existing problems and countermeasures of the application of clustering algorithm in the specialized management of tax sources, and studies and designs the model of the application of clustering algorithm in the specialized management of tax sources. Then, it uses SPSS Clementine20.0 software to study and design the model of the application of clustering algorithm in the specialized management of tax sources. At the same time, the standard model is established according to the process of data mining, and the K-means clustering method is used to analyze 196 companies in the daily consumer industry under WIND classification through the index system selected in this paper. The applicability of clustering algorithm in the application of tax source specialized management is demonstrated.
【学位授予单位】：集美大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F812.42

【参考文献】