Anomaly detection model training method, anomaly detection methods, and load detection devices for household electricity｜Patents Search｜Patents｜Service

2023.10.11

Anomaly detection model training method, anomaly detection methods, and load detection devices for household electricity

United States

Overview

An anomaly detection model training method, an anomaly detection method, and a loading detection device for household electricity consumption are provided, wherein the anomaly detection method is applied to the loading detection device which includes a processing unit and an anomaly detection model. The processing unit performs a training process on the anomaly detection model and executes the anomaly detection method, wherein the anomaly detection method includes: extracting electricity consumption features of a user for obtaining feature data; distributing the feature data into groups according to electricity consumption behaviors, and inputting the groups into a first one-class classification model and a second one-class classification model respectively to generate detection results corresponding to the first one-class classification model and the second one-class classification model; generating an anomaly electricity consumption detection result according to the detection results; distributing an electricity consumption data by time intervals, and comparing the electricity consumption data with historical load data based on the anomaly electricity consumption detection result for outputting anomaly electricity consumption periods.