Unsupervised learning procedure for NILM applications

In a domestic context, NILM applications (NonIntrusive Load Monitoring) allow users to know their electric consumption per appliance without having to install sensors for each appliance in their house. The aim of this paper is to present a new generic method allowing to discover the different appliances present in a house. Results obtained in this paper relies on supervised data obtained by sub-metering at the level of each appliance for 3 real domestic houses.