Quantifying Performance and Uncertainty: A Comprehensive Approach to Time Series Forecasting in a Business Context

Résumé

The aim of this master's thesis is to evaluate time series forecasts. The primary purpose of this work is to offer an insightful perspective on predictive accuracy by taking a new perspective into account: uncertainty. To this end, an exploration of existing metrics in the literature has been carried out, creating a framework for evaluation. In addition, a methodology for quantifying uncertainty, adjusted to a business context, has been developed. This methodology is designed to generate an in-depth analysis that allows uncertainty to be considered. These methods for measuring and creating uncertainty will then be implemented for a business use case, in this case sales forecasts for the Carmeuse company. This application results in an analysis and discussion of the methodologies deployed, highlighting their contributions and future directions.

la date de réponse	28 août 2023
langue originale	Anglais
L'institution diplômante	Universite de Namur
Superviseur	Corentin Burnay (Promoteur)