Abstract
Over the past few years, the number of sensors spread across cities has significantly increased. This led to an exponential growth in data volume, which can only be treated with Big Data techniques. Having such a large amount of generated data, turns possible to apply machine learning techniques more accurately, with the goal of making data predictions over time, finding anomalies, performing classification, among other tasks. This article aims to show the application of machine learning techniques, using a variant of the Recurrent Neural Networks, the Long Short-Term Memory (LSTM), in order to predict city's energy consumptions for the near future. This forecast will support municipal entities decisions, helping them to improve the managing of energy consumptions and budgets.
Recommended Citation
Sério, Francisco; Pinto, Filipe; and Wanzeller, Cristina, "Machine Learning techniques for energy consumption forecasting in Smart Cities scenarios" (2020). CAPSI 2020 Proceedings. 27.
https://aisel.aisnet.org/capsi2020/27