Non-Intrusive Load Monitoring using Deep Learning
Energy Disaggregation done with plain mathematical functions might not cut to the chase in production-grade pipelines, with emphasis on generalizability and superior performance. Modern Deep Learning algorithms have proven to be particularly good architectures in mitigating basic mathematical and Machine Learning implementations. Also, the feature extraction is taken care of by the model itself since we are not handcrafting any features. Since the data size is in the magnitude of hundreds of GBs, it becomes more imperative to go for Deep Learning methodologies for a more robust Energy Disaggregation.