Volume 33, Issue 1, June 2021, Pages 105–121
Joseph Kasiama Ngi-Onkor1, Rostin Mabela Matendo2, Ruffin-Benoît M. Ngoie3, and Jean Jacques Katshitshi4
1 Mathématiques et Info., Fac. Sc., U.P.N., KIN I, RD Congo
2 Dpt. Mathématiques et Info., Fac. Sc., Université de Kinshasa, RD Congo
3 Dpt. Mathématiques et Info., Sec. Sc. Ex., I.S.P. MBANZA-NGUNGU, RD Congo
4 Dpt. A.I.A., Fac. Lettres & Sc. Hum., Université de Kinshasa, RD Congo
Original language: French
Copyright © 2021 ISSR Journals. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
This article suggests a generalized method of discrimination which extends the classical Discriminating Factorial Analysis (FDA) to symbolic objects. This method is based on the adaptation of the classical Bayesian rule of discrimination to symbolic objects. This adaptation is done taking into account various elements, namely: a certain « density » measure on the observation space of symbolic objects; discriminant functions giving an idea of the « similarity » which exists between an observation and the individuals of the formation whole. This rule depends on the formation data and is typically built of the in view the minimization of the overall error rate.The purpose of this study is to solve a capital medical problem. Indeed, several cases of sudden deaths are noted these last years in the whole world and more particularly in our country the Democratic Republic of Congo (RDC), due to Cerebral Vascular Accidents (CVA) or acute coronary events (Heart attacks). The evolution and the prevailence of these cardiovascular diseases present a certain number of real and urgent problems to policy makers and other medical officials. Many undertaken epidemiologic studies these 20 last years led to the identification of the principal Factors of cardiovascular risk (FCVR), opening the way to preventive treatment. It is on the basis of these factors of risk that we designed a tool of decision-making aid medical.This generalized method of discrimination therefore makes it possible to produce decisions concerning whether or not a data point belongs to a predefined class, by using formation sets, from an assigning algorithm of the symbolic objects to classes that we suggest here.
Author Keywords: discrimination method, discriminant factoriel analysis, symbolic objects, density function, bayesian rule, decision rule, discriminant function, formation set, factors of cardiovascular risk.
Joseph Kasiama Ngi-Onkor1, Rostin Mabela Matendo2, Ruffin-Benoît M. Ngoie3, and Jean Jacques Katshitshi4
1 Mathématiques et Info., Fac. Sc., U.P.N., KIN I, RD Congo
2 Dpt. Mathématiques et Info., Fac. Sc., Université de Kinshasa, RD Congo
3 Dpt. Mathématiques et Info., Sec. Sc. Ex., I.S.P. MBANZA-NGUNGU, RD Congo
4 Dpt. A.I.A., Fac. Lettres & Sc. Hum., Université de Kinshasa, RD Congo
Original language: French
Copyright © 2021 ISSR Journals. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
This article suggests a generalized method of discrimination which extends the classical Discriminating Factorial Analysis (FDA) to symbolic objects. This method is based on the adaptation of the classical Bayesian rule of discrimination to symbolic objects. This adaptation is done taking into account various elements, namely: a certain « density » measure on the observation space of symbolic objects; discriminant functions giving an idea of the « similarity » which exists between an observation and the individuals of the formation whole. This rule depends on the formation data and is typically built of the in view the minimization of the overall error rate.The purpose of this study is to solve a capital medical problem. Indeed, several cases of sudden deaths are noted these last years in the whole world and more particularly in our country the Democratic Republic of Congo (RDC), due to Cerebral Vascular Accidents (CVA) or acute coronary events (Heart attacks). The evolution and the prevailence of these cardiovascular diseases present a certain number of real and urgent problems to policy makers and other medical officials. Many undertaken epidemiologic studies these 20 last years led to the identification of the principal Factors of cardiovascular risk (FCVR), opening the way to preventive treatment. It is on the basis of these factors of risk that we designed a tool of decision-making aid medical.This generalized method of discrimination therefore makes it possible to produce decisions concerning whether or not a data point belongs to a predefined class, by using formation sets, from an assigning algorithm of the symbolic objects to classes that we suggest here.
Author Keywords: discrimination method, discriminant factoriel analysis, symbolic objects, density function, bayesian rule, decision rule, discriminant function, formation set, factors of cardiovascular risk.
Abstract: (french)
Cet article propose une méthode généralisée de discrimination qui étend l'Analyse Factorielle Discriminante (AFD) classique aux objets symboliques. Cette méthode est basée sur l’adaptation de la règle bayésienne classique de discrimination aux objets symboliques. Cette adaptation s’effectue en tenant compte de plusieurs éléments, notamment: une certaine mesure de « densité » sur l'espace d'observation des objets symboliques; des fonctions discriminantes donnant une idée de la « similarité » qui existe entre une observation et les individus de l’ensemble de formation. Cette règle dépend des données de formation et est typiquement construite en vue de la minimisation d'un taux d'erreur global.Le but de cette étude est de résoudre un problème médical capital. En effet, plusieurs cas de morts subites sont constatés ces dernières années dans le monde entier et plus particulièrement dans notre pays la République Démocratique du Congo (RDC), dues aux Accidents Vasculaires Cérébraux (AVC) ou aux Evénements coronariens aigus (Crises cardiaques). L’évolution et la prévalence de ces maladies cardiovasculaires présentent un certain nombre de problèmes réels et urgents aux décideurs politiques et autres responsables médicaux. De nombreuses études épidémiologiques menées ces 20 dernières années ont conduit à l’identification des principaux Facteurs de risque cardiovasculaire (FRCV), ouvrant la voie du traitement préventif. C’est sur base de ces facteurs de risque que nous avons conçu un outil d’aide à la décision médicale.Cette méthode généralisée de discrimination permet donc de produire des décisions concernant l'appartenance ou non d'un point de données à une classe prédéfinie, en utilisant des ensembles de formation, à partir d’un Algorithme d'assignation des objets symboliques à des classes que nous proposons ici.
Author Keywords: méthode de discrimination, analyse factorielle discriminante, objets symbolique, fonction de densité, règle bayésienne, règle de décision, fonction discriminante, ensemble de formation, facteurs de risque cardiovasculaire.
How to Cite this Article
Joseph Kasiama Ngi-Onkor, Rostin Mabela Matendo, Ruffin-Benoît M. Ngoie, and Jean Jacques Katshitshi, “Contribution à l’analyse factorielle discriminante des données symboliques: Application aux maladies cardiovasculaires,” International Journal of Innovation and Applied Studies, vol. 33, no. 1, pp. 105–121, June 2021.
