Human Skin Profiling by Physical Skin Biomarkers: A Machine Learning Approach
Personalized skincare requires customization of products, which are adapted in a way to suit patient’s skin profile. The process of customization involves several major steps: (i) determine the information which will be used to successfully distinguish between patient’s skin profiles and the methods to obtain the information, (ii) develop an efficient algorithm which will accurately and efficiently classify the patients according to the information used to distinguish between skin profiles, (iii) determine the correct or most efficient skincare formula to suit the particular skin profile. This study considers the first two steps. First, it was examined whether patient’s physical skin measurements can be used to determine the patient’s skin profile. In the second step, twenty machine learning algorithms, which were selected after initial screening, were employed for the classification task. The study showed that the use of physical skin measurements to distinguish between skin profiles represents a promising option. It was also shown that some of the machine learning algorithms are particularly suitable for classification tasks of this type. Sensitivity of the selected classification algorithms to the location on the skin which is sampled, i.e. affected or unaffected part of skin for volunteers with diabetes mellitus type II and rosacea, is reported.
KeywordsPersonalized skincare Physical skin biomarkers Classification of skin profiles Machine learning algorithms Skin profiling
The present study was partially supported by the FP7 EC Programme, Theme FoF.NMP.2013-6 Mini-factories for customized products using local flexible production, Grant agreement no: 609198.
- 7.Grubinger T, Zeileis A, Pfeiffer K-P (2011) EVTREE: evolutionary learning of globally optimal classification and regression trees in R. Department of Economics (Inst. für Wirtschaftstheorie und Wirtschaftsgeschichte)Google Scholar
- 9.Hsu C-W, Chang C-C, Lin C-J (2003) A practical guide to support vector classificationGoogle Scholar
- 10.Izenman AJ (2013) Linear discriminant analysis. In: Modern multivariate statistical techniques. Springer, pp 237–280Google Scholar
- 11.Kohonen T (1995) Learning vector quantization. In: Self-organizing maps. Springer, pp 175–189Google Scholar
- 18.R Core Team (2014) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://doi.org/10.1017/cbo9781107415324.004
- 19.Ripley BD (2007) Pattern recognition and neural networks. Cambridge university pressGoogle Scholar
- 24.Therneau TM, Atkinson EJ (1997) An introduction to recursive partitioning using the RPART routine. Stats 116:1–52Google Scholar
- 28.Zeevi D, Korem T, Zmora N, Israeli D, Rothschild D, Weinberger A, Ben-Yacov O, Lador D, Avnit-Sagi T, Lotan-Pompan M, Suez J, Mahdi JA, Matot E, Malka G, Kosower N, Rein M, Zilberman-Schapira G, Dohnalová L, Pevsner-Fischer M, Bikovsky R, Halpern Z, Elinav E, Segal E (2015) Personalized nutrition by prediction of glycemic responses. Cell 163:1079–1095. https://doi.org/10.1016/j.cell.2015.11.001CrossRefGoogle Scholar