摘要
为了提高电网企业审计风险管理效果,通过大数据技术挖掘进行电网企业审计风险指标体系构建。首先采用主成分分析法与专家法筛选出电网企业主要的风险因子。其次,引入随机森林算法来构建审计风险预警模型预警,并引入狮群算法来优化参数问题,构建改进的随机森林预警模型。在精确率、召回率测试中,研究模型三种模型表现最好,分别为0.968与0.986,优于支持向量机模型与传统随机森林模型。同时,对于13个风险指标进行预警测试,研究模型预警准确率为96.5%,优于别的模型。由此可见,所提出预警模型整体应用效果更出色,研究内容对电网企业审计风险管理以及智能化发展提供重要的技术支持。
In order to improve the effectiveness of audit risk management in power grid enterprises,a risk indicator system for power grid enterprise audits is constructed through big data technology mining.Firstly,principal component analysis and expert method are used to screen out the main risk factors of power grid enterprises.Secondly,the random forest algorithm is introduced to construct an audit risk warning model,and the lion swarm algorithm is introduced to optimize the parameter problem and construct an improved random forest warning model.In accuracy and recall tests,the three models studied performed the best,with values of 0.968 and 0.986,respectively,outperforming support vector machine models and traditional random forest models.At the same time,warning tests were conducted on 13 risk indicators,and the research model's warning accuracy was 96.5%,which is better than other models.It can be seen that the overall application effect of the proposed early warning model is better,and the research content provides important technical support for audit risk management and intelligent development of power grid enterprises.
作者
严元琪
吴非
Yan Yuanqi;Wu Fei(State Grid Corporation of China Limited,BeiJing 100000)
出处
《现代科学仪器》
2024年第4期208-213,共6页
Modern Scientific Instruments
关键词
随机森林
主数据挖掘
成分分析法
审计风险
预警模型
Random forest
Master data mining
Component analysis method
Audit risks
Early warning model
