Content

The recent years, we at Vitec have been focusing our model development resources in developing AI/Machine Learning models. These constitutes a breaking point in how to forecast energy series which will have impact both on the forecast accuracy as well as the need for administrative tasks concerning the models.

We will start with a brief introduction to machine learning models in general and continue with the implications for the daily forecasting process as well as for the administration of the models. We will describe how we set up  customized model testing environments for each configuration, dedicated for AI model testing. Also, all the steps in the process from testing a new model in the test configuration to using the models in the production environment will be presented. We will also mention the means by which the forecast accuracy of the AI models will be monitored and evaluated on a daily basis.

Finally, some of our results for the AI models so far will be presented, especially for our most recent developed model, which can simultaneously  produce forecasts for a number of series which improves the calculation performance considerably.

How to sign up

As an Educandi member you will get invitations by e-mail.
You can accept the invitation directly or in your calendar. If you haven´t got an invitation please contact marie.ek@vitecsoftware.com.

Presenters

Alexander Nordling
Application Consultant

Christer Modin
Head of Model Development