摘要
针对传统财务报表审计方法存在自动化和智能化水平不高,导致财务报表审计效率低的问题,结合RPA技术的特性,设计一个基于词共现与SOM神经网络的财务报表审计机器人。首先,基于RPA技术对财务报表审计机器人的整体结构进行设计,通过RPA技术实现自动核账、发票识别等功能;然后采用词共现与SOM神经网络实现审计线索的发现;最后通过以上模块实现财务报表审计机器人自动审计。实验结果表明,RPA机器人通过文本聚类,可发现审计中的线索,且线索准确率达85%,高于传统的算法。由此说明,通过RPA技术结合人工智能算法可实现财务审计的智能化,更好地提高财务效率。
In view of the problem that the low automation and intelligence level of traditional financial statement audit methods leads to low efficiency of financial statement audit,combined with the characteristics of RPA technology,a financial statement audit robot based on word co-occurrence and SOM neural network is designed.Firstly,the overall structure of financial statement audit robot is designed based on RPA technology,and automatic audit and invoice recognition are realized through RPA technology;Then,word co-occurrence and SOM neural network are adopted to realize the discovery of audit clues;Finally,the automatic audit of financial statement audit robot is realized through the above modules.The experimental results show that the RPA robot can find the clues in the audit through text clustering,and the clue accuracy is 85%,which is higher than that of the traditional algorithm.Therefore,it shows that RPA technology combined with artificial intelligence algorithm can realize the intelligence of financial audit and better improve the financial efficiency.
作者
薛媛
XUE Yuan(Shaanxi Technical College of Finance&Economics,Xianyang,Shaanxi 712000,China)
出处
《自动化与仪器仪表》
2023年第9期210-214,共5页
Automation & Instrumentation
基金
陕西省教育科学“十三五”规划课题《基于“产教创融合”生产型实训基地的1.5元制实践育人模式探索与研究》(SGH20Y1626)。
关键词
RPA技术
词共现
SOM神经网络
财务审计
发票识别
RPA technology
word co-occurrence
SOM neural network
financial audit
invoice recognition
作者简介
薛媛(1990-),女,陕西西安人,硕士研究生,讲师。
