An AETA Geoacoustic Signal Anomaly Detection Method Based on FindCBLOF.
ICISS(2021)
摘要
The multi-component earthquake monitoring and prediction system AETA has deployed more than 200 devices in China and accumulated a large amount of observation data. How to extract anomalies from AETA geoacoustic signal is a problem worthy of attention. This paper proposes an AETA geoacoustic signal anomaly detection method based on findCBLOF. The AETA geoacoustic data is clustered according to the similarity of data in different time periods. The abnormal score is expressed by CBLOF. The abnormal score is compared with the defined threshold to get the final abnormal value. Through experiments on 10 stations in Sichuan, China, the results show that the precision rate of the method proposed in this paper is 70.8%, the recall rate is 61.8%, and the F1-score reaches 65.8%. Compared with other anomaly detection algorithms, the one based on findCBLOF has effective anomaly extraction capability for AETA geoacoustic data, and has a good correspondence with local earthquakes.
更多查看译文
关键词
findcblof
AI 理解论文
溯源树
样例
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要