Recurrent Neural Networks for Time Series Forecasting

A collection of examples for using RNNs for time series forecasting with Keras. The examples include:

• 1_one_step_univariate.ipynb - model that predicts one step ahead with univariate time series
• 2_one_step_multivariate.ipynb - example of using multivariate time series data
• 3_multi_step_vector_output.ipynb - model that outputs a vector of predictions to forecast multiple steps ahead
• 4_multi_step_encoder_decoder_simple.ipynb - a simple encoder-decoder approach to multi-step forecasting
• 5_multi_step_encoder_decoder_teacher_forcing.ipynb - a more complex encoder-decoder architecture in which the decoder is trained using a teacher forcing approach

The data in all examples is from the GEFCom2014 energy forecasting competition¹. It consists of 3 years of hourly electricity load data from the New England ISO and also includes hourly temperature data. Before running the notebooks, download the data from https://www.dropbox.com/s/pqenrr2mcvl0hk9/GEFCom2014.zip?dl=0 and save it in the data folder.

Prerequisites

You must have the following software and packages installed to run these notebooks:

• Anaconda
• Keras

¹Tao Hong, Pierre Pinson, Shu Fan, Hamidreza Zareipour, Alberto Troccoli and Rob J. Hyndman, "Probabilistic energy forecasting: Global Energy Forecasting Competition 2014 and beyond", International Journal of Forecasting, vol.32, no.3, pp 896-913, July-September, 2016.