Using the Mean Absolute Percentage Error for Regression Models

Arnaud de Myttenaere (1) , Boris Golden (2) , Bénédicte Le Grand (3) , Fabrice Rossi (1)

1 SAMM - Statistique, Analyse et Modélisation Multidisciplinaire (SAmos-Marin Mersenne)
2 Viadeo - Viadeo SA
3 CRI - Centre de Recherche en Informatique de Paris 1

We study in this paper the consequences of using the Mean Absolute Percentage Error (MAPE) as a measure of quality for regression models. We show that finding the best model under the MAPE is equivalent to doing weighted Mean Absolute Error (MAE) regression. We show that universal consistency of Empirical Risk Minimization remains possible using the MAPE instead of the MAE.

Mots clés: Mean Absolute Percentage Error; Regression; Empirical Risk Minimization; Learning Theory; VC-dimension; Covering Numbers

Domaines: Machine Learning [stat.ML]

Arnaud de Myttenaere, Boris Golden, Bénédicte Le Grand, Fabrice Rossi. Using the Mean Absolute Percentage Error for Regression Models. European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN), Apr 2015, Bruges, Belgium. ⟨hal-01162980⟩