Análisis de modelos de redes neuronales artificiales, para un sistema de diagnósticos de migrañas con aura y sin aura
Cargando...
Fecha
2014
Autores
Romero de la Hoz, Zuli
Rúa Ascar, Juan Manuel
Título de la revista
ISSN de la revista
Título del volumen
Editor
Ediciones Universidad Simón Bolívar
Facultad de Ingenierías
Facultad de Ingenierías
Resumen
El estudio se fundamentó en la realización de un análisis comparativo de redes
neuronales artificiales con dominio de aplicación similar, para determinar el modelo y la
arquitectura que mejor clasifique los síntomas de pacientes que presentan diagnósticos de
migrañas con aura y sin aura. De igual forma el resultado del estudio está proporcionando el
conocimiento que puede permitir desarrollar en el futuro, un software que ofrezca al médico
general una herramienta de apoyo, para minimizar los tiempos en la obtención de un
diagnóstico clínico oportuno y acertado, de tal manera que el medico realice una comparación
entre su razonamiento y el resultado que entregue la herramienta.
Los modelos de redes neuronales artificiales implementados en el estudio son: Modelo
Feed forward, Función de Base Radial y LVQ, los cuales han sido objeto de aplicación en
otros estudios como: El diagnóstico de glaucoma, anomalías visuales acromáticas,
insuficiencia cardiaca, detección de cáncer cervical uterino, diagnostico de epilepsia, Estudio
sobre técnicas de análisis de encefalogramas basados en redes neuronales, selección de
personal, aplicación de redes neuronales artificiales a la recuperación de la información,
aplicadas al análisis de datos, entre otros.
Las redes neuronales artificiales como podemos apreciar, son una de las herramientas
inteligentes más implementadas en la solución de diversos problemas donde se requiere
clasificar, analizar y categorizar información. (De Barros Ruano, 2009)
The study was based on the realization of a comparative analysis of artificial neural networks with similar application domain, to determine the model and architecture that best classify the symptoms of patients with diagnoses of migraine with aura and without aura. Similarly the result of the study is to pro vide knowledge that may develop in the future allow a software that provides the general practitioner a support tool to minimize time in obtaining timely and accurate clinical diagnosis, such that the doctor perform a comparison between the reasoning and the tool that delivers results . The models of artificial neural networks implemented in the study are: Model Feed forward, Radial Basis Function and L VQ, which have been applied in other studies such as: The diagnosis of glaucoma, achromatic visual abnormalities, heart failure, cancer detection uterine cervical, diagnosis of epilepsy, Study on EEG analysis techniques based on neural networks, recruitment, application of artificial neural networks to information retrieval, applied to data analysis, among others. Artificial neural networks as we can see, are one of the more intelligent tools implemented in solving various problems where it is required to classify, analyze and categorize information. (De Barros Ruano, 2009)
The study was based on the realization of a comparative analysis of artificial neural networks with similar application domain, to determine the model and architecture that best classify the symptoms of patients with diagnoses of migraine with aura and without aura. Similarly the result of the study is to pro vide knowledge that may develop in the future allow a software that provides the general practitioner a support tool to minimize time in obtaining timely and accurate clinical diagnosis, such that the doctor perform a comparison between the reasoning and the tool that delivers results . The models of artificial neural networks implemented in the study are: Model Feed forward, Radial Basis Function and L VQ, which have been applied in other studies such as: The diagnosis of glaucoma, achromatic visual abnormalities, heart failure, cancer detection uterine cervical, diagnosis of epilepsy, Study on EEG analysis techniques based on neural networks, recruitment, application of artificial neural networks to information retrieval, applied to data analysis, among others. Artificial neural networks as we can see, are one of the more intelligent tools implemented in solving various problems where it is required to classify, analyze and categorize information. (De Barros Ruano, 2009)
Descripción
Palabras clave
Redes Neuronales Artificiales, Migrañas, Diagnósticos, Artificial Neural Networks, Migraine, Diagnostics