基于深度学习的责任审计系统研究

Design of Responsibility Audit System Based on Deep Learning

责任审计对于办事机关事件约束起到积极作用,如何运用技术手段提升审计的效果是业界研究的重点工作。本文结合信息检索(IR)和自然语言处理(NLP)科学的“字典查找”方法来检测单词拼写错误,协助审计团队进行文件分析,实现责任审计的目的,该系统解决在文档发布、报告或审核活动的真实证据之前,必须经过编辑阶段以纠正是否存在错误和不足,系统基于流深度学习算法、字典查找方法有效地根据词汇资源确定单词的正确或错误,证明已开发的字符串匹配方法可以正确、快速地纠正单词书写错误。

Responsible auditing plays an active role in restraining incidents of offices.How to use technical means to improve the effectiveness of auditing is a key task of industry research.This article combines information retrieval(IR)and natural language processing(NLP)scientific"dictionary lookup"methods to detect word spelling errors,assist the audit team in file analysis,and achieve the purpose of responsibility audit.The system solves the problem of document publishing,reporting,or review before the real evidence of the activity,it must go through the editing stage to correct whether there are errors and deficiencies.The system effectively determines the correct or wrong words based on the vocabulary resources,the stream deep learning algorithm and dictionary search method,proves that the developed string matching method can be correct,quickly correct word writing errors.