Multidimensional Feeding of LSTM Networks for Multivariate Prediction of Energy Time Series

We propose a deep learning approach for multi-variate forecasting of energy time series. It is developed by using Long Short-Term Memory deep neural networks so that different related time series, incorporating information of long-term dependencies, can be joined together as a multidimensional input of the deep neural network. The learning scheme can be represented as a stacked LSTM network in which one or more layers are cascaded, feeding their output to the input of the sequent layer. To prove the effectiveness of the approach, it has been tested on real-world problems pertaining to the energy field, where time series prediction is of paramount importance..