Predicting Households’ Short-Term Power Consumption Utilizing LSTM

Due to the growing trend of power consumption, its efficient planning plays a vital role not only in households but also on a national level and beyond. Numerous methods and approaches have been proposed to tackle the complexities associated with power consumption predictions, yet the challenges persist. In this paper, we focus on short-term household power consumption prediction. For that, we briefly present the current techniques that solve this problem, along with their weaknesses - they are usually highly dependent on the nature of the problem and, more precisely, the data on which they are trained. To address this problem, we propose the LSTM model, which was tested on a real-life household power consumption dataset. The results show that the proposed model performs better than the ARIMA method, one of the most common methods for addressing such tasks regarding MAE, MSE, and RMSE metrics, thus providing more accurate and reliable predictions.