Smart Meter Data Analytics Using Decision Trees and Nearest-Neighbours

Smart meters have become increasingly useful and been replacing the traditional electricity meter in the recent past. Smart meters produce a vast amount of data, thereby, making it challenging and complex enough to analyse the data efficiently. Therefore, computational algorithms are necessary to improve the analysis of smart meter data. In this article, algorithms such as decision trees, and nearest-neighbours have been used to analyse smart meter data for load identification and abnormality detection. Data utilized in this study were collected from a real smart meter used for a residential apartment of a metro city. From the results, it was observed that decision trees and nearest-neighbours produced similar results that are accurate for load identification and abnormality detection when compared to results produced with that of the decision trees.