Home Appliances Classification Using Machine Learning Algorithms

Abstract: In present times data driven research is dominating every sector of science and engineering Electricity and power related research has no exception in this regard. Numerous sensors and communication technologies has enabled the researchers to sense and record real time electricity usage scenario. This data are being utilized to develop home energy management systems (HEMS) which can ensure efficient use of electricity. Load monitoring is an important part of HEMS which requires identification or classification of different electrical home appliances from either aggregated or disaggregated electricity consumption data. Home appliance identification can be used in electricity theft detection, occupancy detection. determining equipment condition. This can also be used to put additional information in consumer bill to indicate individual appliance's power consumption and cost so that that appliance can be rescheduled to operate at some other time. Supervised machine learning techniques are extensively used in classification problems. The availability of load signature database supervised classification algorithms are being used by many researchers in home appliance classification. In this work we used very common and popular machine learning  algorithms as decision tree, support vector machine, and random forest to build classification model for appliance recognition. We trained and validated our model using the ASC-F database

• Supervision