Skip to main content


  • Introduction to the open-source Darts library for time series forecasting in Python
  • Get hands-on experience in using traditional statistical-based methods (Exponential
    Smoothing, ARIMA), machine learning models (Linear Regression, LightGBM) and
    deep-learning models (N-BEATS) for time series forecasting
  • Evaluate and compare the performance of the different methods as well as interpret the
    prediction of these models
  • Leverage meta-learning methods to forecast previously unseen time series
  • Practice with well-known datasets M4 and M3 which contains a wide variety of time


09:00 – 09:30 Intro to the open-source Darts library and time series forecasting
09:30 – 10:30 Practice with different methods (Exponential Smoothing, ARIMA, Linear Regression, LightGBM, N-BEATS)
10:30 – 10:45 Coffee Break
10:45 – 12:00 Meta-Learning with N-BEATS and comparison between Meta-Learning and traditional approaches