Margaux Zaffran will defend her Phd “Post-hoc predictive uncertainty quantification: methods with applications to electricity price forecasting” on Tuesday June 25 2024.
Title
Post-hoc predictive uncertainty quantification: methods with applications to electricity price forecasting
Supervisors
Aymeric Dieuleveut (CMAP, Ecole polytechnique) and Julie Josse (Inria, Montpellier), as well as Olivier Féron and Yannig Goude from EDF R&D, as part of a CIFRE contract
The jury will be composed of:
- Pierre Pinson, Professor, Imperial College London (Rapporteur)
- Etienne Roquain, Maître de conférences HdR, Sorbonne Université (Rapporteur)
- Emmanuel Candès, Professor, Stanford (Examiner)
- Florence Forbes, Research Director, Inria, Grenoble (Examiner)
- Éric Moulines, Professor, Ecole polytechnique (Examiner)
- Aaditya Ramdas, Assistant Professor, Carnegie Mellon University (Examiner)
- Aymeric Dieuleveut, Professor, École polytechnique (Thesis supervisor)
- Julie Josse, Research Director, Inria, Montpellier (Thesis supervisor)
- Gilles Blanchard, Professor, Université Paris-Saclay (Invited)
- Olivier Féron, Senior researcher, EDF R&D (Invited)
- Yannig Goude, Senior researcher, EDF R&D (Invited)
Abstract
The surge of more and more powerful statistical learning algorithms offers promising prospects for electricity prices forecasting. However, these methods provide ad hoc forecasts, with no indication of the degree of confidence to be placed in them. To ensure the safe deployment of these predictive models, it is crucial to quantify their predictive uncertainty. This PhD thesis focuses on developing predictive intervals for any underlying algorithm. While motivated by the electrical sector, the methods developed, based on Split Conformal Prediction (SCP), are generic: they can be applied in many sensitive fields.
First, this thesis studies post-hoc predictive uncertainty quantification for time series. The first bottleneck to apply SCP in order to obtain guaranteed probabilistic electricity price forecasting in a post-hoc fashion is the highly non-stationary temporal aspect of electricity prices, breaking the exchangeability assumption. The first contribution proposes a parameter-free algorithm tailored for time series, which is based on theoretically analysing the efficiency of the existing Adaptive Conformal Inference method. The second contribution conducts an extensive application study on novel data set of recent turbulent French spot prices in 2020 and 2021.
Another challenge are missing values (NAs). In a second part, this thesis analyzes the interplay between NAs and predictive uncertainty quantification. The third contribution highlights that NAs induce heteroskedasticity, leading to uneven coverage depending on which features are observed. Two algorithms recovering equalized coverage for any NAs under distributional assumptions on the missigness mechanism are designed. The forth contribution pushes forwards the theoretical analysis to understand precisely which distributional assumptions are unavoidable for theoretical informativeness. It also unifies the previously proposed algorithms into a general framework that demontrastes empirical robustness to violations of the supposed missingness distribution.
Website
mzaffran.github.io
