Energy Consumption Forecasting in Smart Households using Metaheuristic Optimization for Deep Learning
Energy consumption is considered one of the most critical problems in the present era due to its economic and environmental impacts. The prevalence of internet of things (IoT) and its role in communicating smart household devices with the internet makes monitoring and managing power consumption more feasible. On the other hand, the fast progress of machine learning techniques encouraged the emergence of efficient energy consumption prediction models. In this paper, we propose an efficient machine learning approach based on long short-term memory (LSTM) and dipper throated optimization to predict the energy consumption of smart households accurately. To assess the performance of the proposed approach, a dataset containing a set of measurements recorded for four months was used in the conducted experiments. The proposed approach is evaluated using eight evaluation metrics to prove its effectiveness. In addition, six other machine learning models and four optimizers were utilized to confirm the superiority of the proposed approach. The recorded results showed that the proposed optimized LSTM outperforms the other models and achieves 55.6% improvement in the root-mean-square error value when compared to the non-optimized LSTM model.
Keywords - Energy consumption, IoT-based Household, Dipper throated optimization, Long short-term memory, Machine learning models.