A Multi-Scale Adaptive Sliding Window-Based Algorithm for Anomaly Detection in Meter - Generated Metering Data Streams
Hua Li, Zhengrong Wu, Wenqin Chen, Xianzhi Yang · 2024
Conventional anomaly detection algorithms for ammeter metering data stream mainly use the peephole connection to generate LSTM state feature acquisition model, which is vulnerable to the dynamic change of time window, resulting in poor performance indicators. Therefore, a multi-scale adaptive sliding window based anomaly detection algorithm for ammeter metering data stream is proposed. That is, the multi-scale adaptive sliding window is used to process the abnormal time sequence sub segment of the data flow, and the meter measurement data flow anomaly detection classifier is designed, thus generating the meter measurement data flow anomaly detection algorithm. The experimental results show that the designed multi-scale adaptive sliding window based meter measurement data flow anomaly detection algorithm has good performance indicators, reliability and certain application value, and has made a certain contribution to improving meter measurement accuracy and comprehensive measurement reliability.