Thesis of Mathilde Guillemot

Apprentissage par la machine et big data cosmétique en vue de la modélisation mathématique et analyse statistique prédictive sur l’efficacité de formules cosmétiques


In order to address the needs of her customers, the cosmetic industry has to be innovative and must offer new products increasingly efficient and regulatory compliant. The tests that are used to prove their performance on humans, whether instrumental (such as the measurement of relief, firmness or hydration) or clinic, are very expensive.

Faced with these constraints, machine learning technics can provide a preprocessing solution in order to be able to predict the efficacy of cosmetic products, while saving time and money.

This dissertation is aimed to developed predictive models of the cosmetic efficacy based on their formulas, but also from the expected efficacy results.

Advisor: Liming Chen

Defense date: tuesday, december 1, 2020

Mr Nait-Ali A. Professeur(e)Université Paris-Est CréteilRapporteur(e)
Mr Hammami M. Professeur(e)Faculté des Sciences de Sfax TunisieRapporteur(e)
Mr Canu StéphaneProfesseur(e)INSA RouenPrésident(e)
Mme Godin C.Ingénieur(e) de rechercheMINATEC - CEA/LETIExaminateur​(trice)
Mme Heusele C.DocteurResponsable de l'Open Innovation - LVMH RechercheCo-encadrant(e)
Mr Korichi R.DocteurResponsable de la recherche in vivo et de l'évaluation des performances de produits - LVMH RechercheCo-encadrant(e)
Mr Chen LimingProfesseur(e)LIRIS - ECLDirecteur(trice) de thèse