Exploration of Anomalous Tracing of Records in Smart Home Energy Consumption Dataset

Abstract

Smart meter captures the energy consumption data at predefined rates and stores it in a file format specified by the utilities. The data of a day should be properly traced into respective file with correct file naming. Any deviation in this process is called anomalous tracing, which affects the accuracy of analytics. Usually, the anomalous tracing is due to the malfunctioning of the metering infrastructure, network congestion, etc. Hence, it is required to identify such tracing to enhance the dataset quality. Thus, this paper proposes a 3-step (extraction, comparison, identification) approach to explore the anomalous tracing in energy consumption dataset. The extraction-step retrieves the date information from the file and file name, which are compared in comparison-step. Further, the irregularities are observed in identification-step. Tracebase dataset is used for the execution, thereby observing that the proposed approach has successfully identified anomalous tracing (irregularity in tracing records and inconsistency in file names).