| Title |
An Experiment and Operational Scenario Design for Anomaly Period Detection using SAPEF-AE |
| Authors |
노태헌(Tae Heon No) ; 이민구(Min Goo Lee) ; 박용국(Yong Kuk Park) ; 정경권(Kyung Kwon Jung) |
| DOI |
https://doi.org/10.5573/ieie.2024.61.6.103 |
| Keywords |
IoT; Anomaly detection; Statistical analysis; AutoEncoder; Operational scenario |
| Abstract |
In the Internet of Things(IoT) field, anomaly detection systems are emerging as important tasks to cope with data quality management threats arising from the unstable interaction environment between sensors and external nodes. Various studies using real-time monitoring, statistical analysis, and AI models are being conducted here, but it is difficult to present smooth experiments and application plans due to limited manpower and time. Accordingly, in this paper, the anomaly period detection model, SAPEF-AE was designed, Statistical-based Period Filtering(SAPEF) and AutoEncoder designed based on the results derived from communication period analysis are integrated, and the problem is overcome by improving the shortcomings and maximizing the advantages of statistical analysis and AI models. The proposed model verified the accuracy and reliability through the meaningful accuracy, precision, and recall evaluated in the experiment conducted based on the actual residence information data in unstable and stable environments collected by the sensor, and a practical application plan is suggested in the operational scenario using SAPEF-AE. |