Implementing Scikit-Learn libralies for developing a Non-Intrusive Load Monitoring program

Abstract

In recent years, load monitoring plays an important role in energy saving and smart grid development. However, detailed monitoring of
individual loads requires a huge number of measuring devices, leading to a lot of difficulties in investment, development and management.
To solve this problem, non-intrusive load monitoring technique which apply machine learning algorithms allows to identify individual loads
consumptions base on total consumption data, thereby significantly reducing the number of measuring devices and investment costs. In this
paper, through the application of machine learning algorithms, we analyzed the energy consumption dataset of an apartment with total and
break-down consumption data. The result point outs difficulties and potentials of applied machine learning in the energy disaggregation.