Assessment Metrics for Unsupervised Non-intrusive Load Disaggregation Learning Algorithms

摘要: Non-intrusive appliance load monitoring (NILM) is the process for disaggregating total electricity consumption into its contributing appliances. Unsupervised NILM algorithm an attractive method as need data annotation can be eliminated. In order to evaluate performance of unsupervised learning algorithm, most research work evaluates using accuracy type metrics. These assessment metrics not only identify disaggregation error, but also distinguish error each other single appliance. better quantify nature we propose two metrics: and a this paper, methods these The experiment results show that our algorithms effectively.