Global Classification (global + classification)

Distribution by Scientific Domains


Selected Abstracts


Global classification of human facial healthy skin using PLS discriminant analysis and clustering analysis

INTERNATIONAL JOURNAL OF COSMETIC SCIENCE, Issue 2 2001
C Guinot
Synopsis Today's classifications of healthy skin are predominantly based on a very limited number of skin characteristics, such as skin oiliness or susceptibility to sun exposure. The aim of the present analysis was to set up a global classification of healthy facial skin, using mathematical models. This classification is based on clinical, biophysical skin characteristics and self-reported information related to the skin, as well as the results of a theoretical skin classification assessed separately for the frontal and the malar zones of the face. In order to maximize the predictive power of the models with a minimum of variables, the Partial Least Square (PLS) discriminant analysis method was used. The resulting PLS components were subjected to clustering analyses to identify the plausible number of clusters and to group the individuals according to their proximities. Using this approach, four PLS components could be constructed and six clusters were found relevant. So, from the 36 hypothetical combinations of the theoretical skin types classification, we tended to a strengthened six classes proposal. Our data suggest that the association of the PLS discriminant analysis and the clustering methods leads to a valid and simple way to classify healthy human skin and represents a potentially useful tool for cosmetic and dermatological research. Résumé Les classifications actuelles qui définissent une peau saine sont fondées principalement sur un nombre très limité de caractéristiques cutanées telles que l'aspect gras de la peau ou sa sensibilité au soleil. Cette analyse a pour but d'établir une classification globale de la peau humaine saine du visage à l'aide de modèles mathématiques. Une recherche de typologie a été effectuée à partir des caractéristiques cliniques et biophysiques de la peau du visage des individus tout en tenant compte d'une classification théorique, reflet de l'expertise des dermatologues, appréciée sur les zones frontale et malaire du visage. Pour accroître l'efficacité prédictive des modèles avec un minimum de variables, la méthode d'analyse discriminante PLS (Partial Least Square) a été utilisée. Des méthodes de classification ont été appliquées aux composantes PLS obtenues afin de déterminer le nombre le plus vraisemblable de classes et pour regrouper les sujets selon leurs proximités. Grâce à cette approche, quatre composantes PLS ont pu être construites et six classes se sont avérées pertinentes. Ce travail a abouti à une proposition de classification à six classes plus vraisemblable et acceptable que les 36 combinaisons possibles de la classification théorique proposée. Nos données suggèrent que l'association de l'analyse discriminante PLS aux méthodes de classification permet d'obtenir de façon simple et appropriée une classification de la peau humaine saine et représente un outil potentiel utile dans le domaine de la recherche en cosmétologie et en dermatologie. [source]


Establishment of grading criteria for acne severity

THE JOURNAL OF DERMATOLOGY, Issue 5 2008
Nobukazu HAYASHI
ABSTRACT For the epidemiological surveys and evaluations of therapy, it is essential to evaluate the severity of diseases. There are several reported methods of assessment for acne severity including lesion counting, comparison of the patient's to a photographic standard and comparison of the patient's to a text description. But all of these are based on opinions of specialists. In this study, we attempted to make an evidence-based grading criteria for acne severity, which was expected to yield consents from most dermatologists. The dermatologists consulted classified the global severity of acne patients without any standard and then counted the numbers of eruptions. Three independent expert dermatologists graded the photographs of these patients. We compared the verdicts of the consulted dermatologist and three experienced dermatologists, and analyzed the relationships between these classifications and numbers of eruptions. Our results showed that most of the dermatologists have similar latent recognitions of acne severity. We selected representative photographs as standards, which would contribute to making adjustments for judgments. Global classifications of dermatologists correlated with numbers of inflammatory eruptions (papules plus pustules), but did not with numbers of comedones. The appropriate divisions of inflammatory eruptions of half of the face to decide classifications were: 0,5, "mild"; 6,20, "moderate"; 21,50, "severe"; and more than 50, "very severe". [source]


Application of DA-preconditioned FINN for electric power system fault detection

ELECTRICAL ENGINEERING IN JAPAN, Issue 2 2009
Tadahiro Itagaki
Abstract This paper proposes a hybrid method of deterministic annealing (DA) and fuzzy inference neural network (FINN) for electric power system fault detection. It extracts features of input data with two-staged precondition of fast Fourier transform (FFT) and DA. FFT is useful for extracting the features of fault currents while DA plays a key role in classifying input data into clusters in a sense of global classification. FINN is a more accurate estimation model than the conventional artificial neural networks (ANNs). The proposed method is successfully applied to data obtained by the Tokyo Electric Power Company (TEPCO) power simulator. © 2008 Wiley Periodicals, Inc. Electr Eng Jpn, 166(2): 39, 46, 2009; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/eej.20497 [source]


Global classification of human facial healthy skin using PLS discriminant analysis and clustering analysis

INTERNATIONAL JOURNAL OF COSMETIC SCIENCE, Issue 2 2001
C Guinot
Synopsis Today's classifications of healthy skin are predominantly based on a very limited number of skin characteristics, such as skin oiliness or susceptibility to sun exposure. The aim of the present analysis was to set up a global classification of healthy facial skin, using mathematical models. This classification is based on clinical, biophysical skin characteristics and self-reported information related to the skin, as well as the results of a theoretical skin classification assessed separately for the frontal and the malar zones of the face. In order to maximize the predictive power of the models with a minimum of variables, the Partial Least Square (PLS) discriminant analysis method was used. The resulting PLS components were subjected to clustering analyses to identify the plausible number of clusters and to group the individuals according to their proximities. Using this approach, four PLS components could be constructed and six clusters were found relevant. So, from the 36 hypothetical combinations of the theoretical skin types classification, we tended to a strengthened six classes proposal. Our data suggest that the association of the PLS discriminant analysis and the clustering methods leads to a valid and simple way to classify healthy human skin and represents a potentially useful tool for cosmetic and dermatological research. Résumé Les classifications actuelles qui définissent une peau saine sont fondées principalement sur un nombre très limité de caractéristiques cutanées telles que l'aspect gras de la peau ou sa sensibilité au soleil. Cette analyse a pour but d'établir une classification globale de la peau humaine saine du visage à l'aide de modèles mathématiques. Une recherche de typologie a été effectuée à partir des caractéristiques cliniques et biophysiques de la peau du visage des individus tout en tenant compte d'une classification théorique, reflet de l'expertise des dermatologues, appréciée sur les zones frontale et malaire du visage. Pour accroître l'efficacité prédictive des modèles avec un minimum de variables, la méthode d'analyse discriminante PLS (Partial Least Square) a été utilisée. Des méthodes de classification ont été appliquées aux composantes PLS obtenues afin de déterminer le nombre le plus vraisemblable de classes et pour regrouper les sujets selon leurs proximités. Grâce à cette approche, quatre composantes PLS ont pu être construites et six classes se sont avérées pertinentes. Ce travail a abouti à une proposition de classification à six classes plus vraisemblable et acceptable que les 36 combinaisons possibles de la classification théorique proposée. Nos données suggèrent que l'association de l'analyse discriminante PLS aux méthodes de classification permet d'obtenir de façon simple et appropriée une classification de la peau humaine saine et représente un outil potentiel utile dans le domaine de la recherche en cosmétologie et en dermatologie. [source]