财务共享服务模式下企业内部智能审计路径研究

发布时间：2018-02-12 10:15

本文关键词： 财务共享服务模式智能审计大数据深度学习　出处：《重庆理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：大型集团对集中管理和数据共享的强烈需求,使得财务共享服务中心在跨国公司和国内大型集团企业中掀起了应用的热潮。财务共享服务模式和传统财务管理模式之间有较大区别,管理模式的改变给内部审计带来了全方位的影响。面对财务共享服务模式,内部审计部门应该如何有条不紊地开展内部审计工作?如何更高效率且高质量完成内部审计数据分析工作?这些都是亟需解决的理论和现实问题。本文运用会计学、审计学、财务管理、计算机应用技术等学科理论,以信息论、系统论作为理论基础和思维向导,充分发挥财务共享服务中心的优势,结合大数据的特点和人工智能技术,构建了财务共享服务模式下企业内部智能审计路径。并开发了内部审计数据分析系统。首先,本文从四个步骤,即建设智能审计信息系统、内部审计人员的组织管理、智能审计实施程序、获取审计证据,一步一步构建出财务共享服务模式下企业内部智能审计路径。在智能审计实施程序中,提出通过智能审计分析进行审计预警,发现审计疑点;并围绕审计疑点采用MECE分析法确定审计重点;再通过审计知识库提供的审计方案,利用非现场审计为主和现场审计为辅的取证模式及时地获取充分、可靠、相关的审计证据。其次,由于利用人工智能技术是实现智能审计的重点,本文把智能审计路径中利用智能数据分析发现审计疑点作为单独一章进行重点讨论。在此部分,在智能审计分析中采用对比分析法,通过从企业内外部的大数据中挖掘出可靠且具有参考价值的比较基准值,与被审单位值对比分析,进行审计预警,从而发现审计疑点。同时,由于财务共享服务模式下的审计数据具备大数据“4V”((Volume、Variety、Velocity、Veracity)特点,在审计数据集市里使用人工操作的方式调用数据进行分析,会造成调用数据不精准且审计效率低的局面。因此,本文通过深度学习网络构建了智能审计数据分析模型,采用自动编码器的原理使数据自学习,主动识别和提取非结构化数据的数字和文本特征,再通过训练数据与选取审计项目下的评价指标进行一一对应,使得审计数据智能匹配到审计项目下进行数据的对比分析,提高审计效率和质量。最后,基于理论部分的分析,利用MATLAB 2015a开发平台、SQL Anywhere9.0数据库,设计并开发了内部审计数据分析系统,同时结合案例的形式印证了智能审计数据分析模型的可行性。
[Abstract]:The strong need for centralized management and data sharing in large groups, The financial sharing service center has set off an upsurge of application in multinational corporations and domestic large group enterprises. There is a great difference between the financial sharing service mode and the traditional financial management mode. The change of the management mode has brought the omnidirectional influence to the internal audit. Facing the financial sharing service mode, how should the internal audit department carry out the internal audit work methodically? How to complete internal audit data analysis efficiently and qualitatively? These are theoretical and practical problems that need to be solved. This paper uses the theories of accounting, auditing, financial management, computer application technology and so on, and takes information theory and system theory as the theoretical basis and thinking guide. This paper gives full play to the advantages of financial sharing service center, combines big data's characteristics and artificial intelligence technology, constructs the internal intelligent audit path of enterprise under the mode of financial sharing service, and develops the internal audit data analysis system. This article from four steps, namely, constructs the intelligent audit information system, the internal auditor organization management, the intelligent audit implementation procedure, obtains the audit evidence, In the implementation program of intelligent audit, it is put forward that the audit warning is carried out through intelligent audit analysis, and the audit doubt is found. MECE analysis method is used to determine the audit focus around the audit doubt points, and then through the audit plan provided by the audit knowledge base, the evidence collection mode of off-site audit and on-the-spot audit is used to obtain sufficient and reliable evidence in time. Secondly, because the use of artificial intelligence technology is the focus of intelligent audit, this paper focuses on the use of intelligent data analysis to find audit suspects in the intelligent audit path as a separate chapter. In the intelligent audit analysis, by using the comparative analysis method, through excavating the reliable and reference reference value of the comparison benchmark value from the inside and outside of the enterprise big data, comparing and analyzing with the value of the audited unit, carrying on the audit early warning, At the same time, because the audit data under the financial sharing service mode has the characteristics of big data "4V" Volumey Varietyn Velocity verity, the audit data is transferred in the audit data set market by manual operation for analysis. Therefore, this paper constructs the intelligent audit data analysis model through the deep learning network, and adopts the principle of automatic encoder to make the data self-learning. The digital and text features of the unstructured data are identified and extracted actively, and then the training data and the evaluation index under the audit item are matched one by one, so that the audit data can be intelligently matched to carry out the comparative analysis of the data under the audit project. Finally, based on the theoretical analysis, the internal audit data analysis system is designed and developed based on the MATLAB 2015a development platform. At the same time, the feasibility of intelligent audit data analysis model is confirmed by case study.
【学位授予单位】：重庆理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F626.115;F239.4

【参考文献】