Abstract:
The escalating integration of smart homes with smart grids underscores the critical need for precise and timely predictions
of energy consumption, essential for optimizing resource allocation and bolstering overall energy efficiency. This research work
pioneers an innovative approach to enhance energy consumption predictions within smart homes by seamlessly integrating the robust
time series forecasting capabilities of the Prophet algorithm with adaptive optimization techniques – ADAM (Adaptive Moment
Estimation), SGD, ADAGRAD, and RMSPROP. Prophet's inherent proficiency in handling daily patterns and seasonality is further
amplified by the adaptability conferred by optimization algorithms, addressing the intricate dynamics of non-linear patterns inherent
in smart home energy consumption. Utilizing the extensive Pecan dataset, encompassing historical energy consumption of various
appliances in a smart home, the proposed hybridized model undergoes rigorous evaluation against traditional Prophet and baseline
models. Metrics such as Mean Absolute Percentage Error (MAPE), Mean Absolute Error (MAE), and Root Mean Squared Error
(RMSE) serve as comprehensive benchmarks for assessing the model's performance. The hybridized model demonstrates a notable
enhancement in accuracy and efficiency in predicting energy consumption, marking a substantial contribution to the ongoing
evolution of energy management practices within smart homes connected to smart grids. As smart homes continue their trajectory of
evolution, the primary aim of this research is to foster sustainable energy practices and optimize resource utilization, aligning with
the ethos of smart living.