Scholarlink

<p>This research introduces an audit multidimensional temporal risk prediction model that integrates variational autoencoders, gated recurrent units, and a position-aware mechanism, to enhance the efficacy of audit risk detection. The model introduces a dual-branch module of temporal attention and position awareness into an improved variational autoencoder-gated recurrent unit structure. It captures multi-scale risk evolution features through sliding window segmentation. In the experiment, the proposed model achieves F1 scores of 97.21% and 96.57% on two types of public datasets, respectively, with a detection latency reduced to approximately 60&#xa0;ms. In the noise robustness experiment, the false negative rate is controlled within 6.52% under 30% noise interference, and the statistical test <i>p</i> value is less than 0.05. The method has broad application potential in intelligent financial auditing, internal control assessment, and enterprise risk management.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Intelligent auditing and risk management based on encoding/decoding structure and gated recurrent units

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月24日

查看全文
  • Jing Chen

摘要

This research introduces an audit multidimensional temporal risk prediction model that integrates variational autoencoders, gated recurrent units, and a position-aware mechanism, to enhance the efficacy of audit risk detection. The model introduces a dual-branch module of temporal attention and position awareness into an improved variational autoencoder-gated recurrent unit structure. It captures multi-scale risk evolution features through sliding window segmentation. In the experiment, the proposed model achieves F1 scores of 97.21% and 96.57% on two types of public datasets, respectively, with a detection latency reduced to approximately 60 ms. In the noise robustness experiment, the false negative rate is controlled within 6.52% under 30% noise interference, and the statistical test p value is less than 0.05. The method has broad application potential in intelligent financial auditing, internal control assessment, and enterprise risk management.