Analytical Approach to Exploring the Missing Data Behavior in Smart Home Energy Consumption Dataset

Smart homes are considered to be the subset of smart grids that have gained widespread popularity and significance in the present energy sector. These homes are usually equipped with different kinds of sensors that communicate between appliances and the metering infrastructure to monitor and trace the energy consumption details. The smart meters trace the energy consumption data continuously or in a period of intervals as required. Sometimes, these traces will be missed due to errors in communication channels, an unexpected breakdown of networks, malfunctioning of smart meters, etc. This missingness greatly impacts smart home operations such as load estimation and management, energy pricing, optimizing assets, planning, decision making, etc. Moreover, to implement a suitable precautionary measure to eliminate missing of data traces, it is required to understand the past behavior of the data anomalies. Hence, it is essential to comprehend the behavior of missing data in the smart home energy consumption dataset. In this regard, this paper proposes an analytical approach to detect and quantify the missing data instants in all days for all appliances. Using this quantification, the behavior of missing data anomalies is analyzed during the day. For the analysis, a practical smart home energy consumption dataset ‘Tracebase’ is considered. Initially, the existence and the count of missing instants are computed. From this, the appliance ‘MicrowaveOven’ is considered for further analysis as it comprises the highest count of missing instants (۸۴۷۴۰) in a day when compared to all other appliances. Finally, the proposed analysis reveals that the large number of missing instants is occurring during the daylight period of a day.

کلیدواژه ها:

behavior analysis ، Data Analysis ، Energy consumption data ، Missing data anomalies ، Smart Homes ، Smart meter data