Forecasting
Regression Tree Methods
Itron will continue with virtual forecasting events again this year. This includes the Annual Energy Forecasting Meeting, also known as the EFG meeting (April 13-15), the ISO Forecasting Summit (May and September), and the Brown Bag Webinars (6 webinars, including one for Europe). We are also adding a new virtual training workshop in June.
The first of the virtual events will be a free brown bag webinar on Regression Tree Methods on Tuesday, Feb. 8 at 12 p.m. PST by Dr. J. Stuart McMenamin who will provide an overview of three methods: Regression Tree, Gradient Boosting and Random Forest. These three nonparametric methods are used to model daily energy use with multiple explanatory factors. The webinar covers the mechanics of the methods, including determination of optimal settings and the use of bootstrap methods to approximate prediction intervals. Out of sample cross validation is used to compare the accuracy of these methods to parametric regression and neural network models.
Other brown bag webinars include:
- Short-term Forecast Stability vs. Accuracy – Controlling the Sun in Europe
- Short-term Forecast Stability vs. Accuracy – Controlling the Sun in North America
- Modeling Climate Change
- 2022 Forecast Accuracy Benchmarking Survey & Energy Trends
- Integrating Distributed Energy Resources (DERs) into Long-term Circuit-level Forecasts