Weather-Driven Model for Predicting Energy Consumption in Future IoT-based Energy Metering Systems

Abstract

The examination of weather's influence on energy consumption originates from the inception of contemporary energy systems. Historically, energy demand was predominantly assessed using seasonal fluctuations and historical consumption data. The digital revolution catalysed the significance of incorporating weather data into energy assessments. Initial research employed fundamental statistical models to establish a correlation between meteorological patterns and energy consumption. Nonetheless, the advent of machine learning methodologies in the past twenty years has transformed this domain. The employment of decision trees, random forests, and neural networks has allowed academics to develop highly precise predictive models. This research utilizes historical developments and advanced technology to investigate the correlation between weather patterns and energy usage in smart homes, thereby advancing the progress of energy-efficient technologies and practices. This research seeks to examine the complex interaction between weather patterns and energy use in smart homes using regression analysis. In addition, it utilizes machine learning approaches to investigate predictive models that analyze the influence of meteorological variables on overall energy demand in these contexts. The proposed methodology employs decision tree and random forest regression techniques, yielding significant insights into energy usage trends across different weather situations.