NILM Algorithms General Comparison and Test of Adaptability
Arthur Pasquet · LA Referencia (Red Federada de Repositorios Institucionales de Publicaciones Científicas) · 2020
NILM field is a hot spot in university and companies research due to the great advantages it \t\t\t\t provides and its importance to reduce energy consumption within the households particularly. \t\t\t\t This thesis allows a comparison between Benchmark and State-of-Art algorithms over various \t\t\t\t datasets from different domains and measured by 12 metrics. It shows that the efficiency of an \t\t\t\t algorithm depends very much on the metric used to measure it. \t\t\t\t As a result, it is observed that algorithms using Deep Learning are generally superior to the \t\t\t\t others, however it is not easy to rank them. The Transfer Learning tried between European \t\t\t\t datasets underlines an encouraging lead, but on the contrary between American dataset it seems \t\t\t\t unproductive. \t\t\t\t This thesis carries out also the first multi-source Transfer Learning in the NILM field, concluding \t\t\t\t the need of further experimentation to prove its relevancy