An easy step-by-step guide to getting started with Neural Networks for Time Series Forecasting
—
Forecasting multiple time series can quickly become a complicated task; traditional approaches either require a separate model per series (i.e. SARIMA) or that all series are correlated (i.e. VARMA). Neural Networks offer a flexible approach that enables multi-series forecasts with a single model regardless of series correlation.
Additionally, this approach allows exogenous variables to be easily incorporated and can forecast multiple timesteps into the future resulting in a powerful general solution that performs well in a wide variety of cases.
In this article, we’ll show how to perform the data windowing required to transform our data from a time series to supervised learning format for both a univariate and multivariate time series. Once our data has been transformed we’ll show how to train both a Deep Neural Network and LSTM to make multivariate forecasts.
Examining Our Data
We’ll be working with a dataset capturing daily mean temperature and humidity in Delhi India between 2013 and 2016. This data is available on Kaggle and is licensed for usage under the CC0: Public Domain making it ideal…