Load monitoring is the key part of demand-side management in power grid. Non-intrusive load monitoring (NILM)is the most promising method in the field. Recently, dictionary-based method has been proven to be effective for NILM.However, existing approaches are all shallow models, which learn only one layer of dictionary representation. As theresult, the representation ability of existing approaches is limited. In order to learn more robust and accurate loadfeatures, this paper proposes a novel approach based on deep dictionary learning. Extensive experiments on REDD datasetverify the effectiveness of our method. The experimental results also show the advantage of deep models against shallowones.
NETL
NSTL
万方数据
