Yannig Goude from EDF (France) gave a seminar at LSE about his research on industrial forecasting.

The talk was very well received and followed by lively discussions.  Given recent technological developments in renewable energy, electric vehicles and smart meters, energy systems are increasingly complex and challenging to model.  Stakeholders are developing smart power grids, including advanced communication networks and associated optimisation and forecasting tools. Efficient use of the smart grid requires advanced data analytics to improve forecasting at different geographical scales.  Yannig’s talk, entitled, “Machine Learning Methods for Electricity Load Forecasting: Contributions and Perspectives“, presented recent developments in online learning and probabilistic forecasting at EDF for this challenge.