Sistema inteligente para pronóstico de supervivencia de paciente con trasplante renal
Fecha
2013-06-24
Autores
Pérez Díaz, Grettel
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central "Marta Abreu" de la Villas
Resumen
Cuba realiza trasplantes renales desde hace más de 34 años, siendo uno de los primeros países de nuestro continente en esta experiencia, el análisis histórico-lógico en torno al tema del trasplante renal conduce a inferir que persiste insuficiencia a la hora de pronosticar el grado de supervivencia de los pacientes trasplantados de riñón. Atendiendo a esta problemática, se necesita automatizar dicho pronóstico teniendo en cuenta las características más significativas de los pacientes y considerando los atributos de los donantes para ayudar a la toma de decisiones basándose en estos parámetros.
Se obtiene un sistema basado en el conocimiento híbrido para la predicción del tiempo de supervivencia del injerto renal de pacientes del Hospital Universitario “Dr. Arnaldo Milián Castro”. Este se desarrolla a partir de la edición de una base de casos obtenida como resultado de la ingeniería de conocimiento, utilizando WEKA se determinan los métodos de aprendizaje que mejores resultados generan en el pronóstico del rasgo objetivo que es continuo y representa tiempo de supervivencia del injerto.
KITS se implementa utilizando java con una interfaz que satisface los requerimientos del usuario. Una red neuronal tipo multicapa pronostica el tiempo de supervivencia y un módulo basado en k-NN explica sobre la base de los casos más similares al pronóstico el porqué de la solución. KITS se encuentra en etapa de evaluación por parte de los expertos en el dominio.
Cuba carries out renal transplants for more than 34 years, being one of the first countries of our continent in this experience, the historical-logical analysis around the topic of the renal transplant leads to infer that inadequacy persists when predicting the grade of the transplanted patients' of kidney survival, for what one has the necessity to obtain a system that is able to offer this presage, keeping in mind the most significant characteristics in the patients, besides helping to the taking of decisions being based on these parameters. A system is obtained based on the hybrid knowledge for the prediction of the time of survival of the renal implant of patient of the University Hospital "Dr. Arnaldo Milián Castro". This it is developed starting from the edition of a base of cases obtained as a result of the engineering of knowledge, using WEKA the learning methods is determined that better results generate in the presage of the objective feature that is continuous and it represents time of survival of the implant. KITS is implemented using java with an interface that satisfies the user's requirements. A net neuronal type multicapa predicts the time of survival and a module based on k-NN explains on the base from the most similar cases to the presage the reason of the solution. KITS is in evaluation stage on the part of the experts in the domain.
Cuba carries out renal transplants for more than 34 years, being one of the first countries of our continent in this experience, the historical-logical analysis around the topic of the renal transplant leads to infer that inadequacy persists when predicting the grade of the transplanted patients' of kidney survival, for what one has the necessity to obtain a system that is able to offer this presage, keeping in mind the most significant characteristics in the patients, besides helping to the taking of decisions being based on these parameters. A system is obtained based on the hybrid knowledge for the prediction of the time of survival of the renal implant of patient of the University Hospital "Dr. Arnaldo Milián Castro". This it is developed starting from the edition of a base of cases obtained as a result of the engineering of knowledge, using WEKA the learning methods is determined that better results generate in the presage of the objective feature that is continuous and it represents time of survival of the implant. KITS is implemented using java with an interface that satisfies the user's requirements. A net neuronal type multicapa predicts the time of survival and a module based on k-NN explains on the base from the most similar cases to the presage the reason of the solution. KITS is in evaluation stage on the part of the experts in the domain.
Descripción
Palabras clave
Sistema Informático, Sistema de Gestión, Sistemas de Información