| 杨春霞,杨士逸,吴一尘,谢陈磊.基于TranAD-FD的多变量时间序列异常检测方法[J].安徽建筑大学学报,2026,34(2):43-50 |
| 基于TranAD-FD的多变量时间序列异常检测方法 |
| A Multivariate Time Series Anomaly Detection Method Based on TranAD-FD |
| |
| DOI： |
| 中文关键词: 特征融合 异常检测 频域 多变量时间序列 |
| 英文关键词: feature fusion anomaly detection frequency domain multivariate time series |
| 基金项目:国家重点研发计划项目（2024YFC3808100） |
| 作者 | 单位 | | 杨春霞 | School of Electronic and Information Engineering,Anhui Jianzhu University,Hefei 230601,China；School of Automation,Nanjing University of Information Science and Technology,Nanjing 210044,China | | 杨士逸 | School of Electronic and Information Engineering,Anhui Jianzhu University,Hefei 230601,China | | 吴一尘 | School of Electronic and Information Engineering,Anhui Jianzhu University,Hefei 230601,China | | 谢陈磊 | School of Electronic and Information Engineering,Anhui Jianzhu University,Hefei 230601,China |
|
| 摘要点击次数: 164 |
| 全文下载次数: 0 |
| 中文摘要: |
| 针对标签缺失时，无监督学习难以辨识多变量时间序列数据中正常与异常数据的问题，提出一种基于TranAD-FD 架构的多变量时间序列异常检测方法。结合 FD 模块与 TranAD，能够有效地避免捕获短期时间趋势。采取滑动窗口划分时间序列，特征融合技术融合多个时间窗的特征，并通过多头自注意机制加权融合，可以增强模型的性能和泛化能力。实验结果表明，该方法不仅具备较高的训练效率与良好的稳定性，而且精确率、召回率、F1值等核心异常检测性能指标的表现也优于当前已有的生成式异常检测模型。 |
| 英文摘要: |
| To address the issue that unsupervised learning struggles to distinguish normal from abnormal data in multivariate time series when labels are missing, a multivariate time series anomaly detection method based on the TranAD-FD architecture is proposed. By combining the FD module with TranAD, the method can effectively avoid capturing short-term time trends. Time series are divided into windows using a sliding window approach, and feature fusion technology is employed to integrate features from multiple time windows and enhance the performance and generalization ability of the model, which not only has high training efficiency and good stability, but also outperforms the existing generative anomaly detection models in core anomaly detection performance indicators such as precision rate, recall rate and F1 value. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
