Anomalies detection for smart-home energy forecasting using moving average

In the past few years, increase in relation between physical and digital world over internet was witnessed. Even though applications can enhance smart home systems, it is still early stages challenges field of things (IoT). An extreme level data quality (DQ) system management essential to produce a meaningful vision. However, most energy has no straightforward process removing abnormal data. Hence, research aims propose validate model anomaly detection for power consumption real-time. The moving average (MA) approach identifies removes results obtained from forecasting time series auto regressive integrated (ARIMA) demonstrated that proposed heuristics effectively enhanced usage forecasting. selection optimum parameter values MA comprehended time-series error minimization by comparing mean squared (MSE). These outcomes proved effectiveness existing technique precision choice appropriate. Therefore, method route cleaned sequence streams real-time environment, which valuable spotting anomalies eliminating enhancing series.