Publication: Problemas de optimización en aprendizaje automático con aplicaciones biomédicas
Loading...
Official URL
Full text at PDC
Publication Date
2023-06-17
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
En este trabajo vamos a abordar el diseño de marcos matemáticos que ayuden en el diagnóstico, tratamiento y posible pronóstico de enfermedades. Nos centraremos en un modelo para el cáncer de próstata. Primero, plantearemos el problema de optimización desde dos estrategias diferentes:
La estrategia determinista: En ella plantearemos el problema de optimización que correspondería al tratamiento del paciente, utilizando como restricción un sistema de EDOS.
La estrategia de aprendizaje: En ella hacemos el mismo planteamiento, pero esta vez, como restricción, utilizaremos una red neuronal previamente ajustada a una colección de datos preseleccionados.
Finalmente, procederemos a resolverlo utilizando los métodos correspondientes a cada uno de los planteamientos propuestos.
In this work we are going to dealt with the design of mathematical frameworks useful in the diagnosis, treatment and possible prognosis of diseases. We will focus on a model for prostate cancer. First, we will pose the optimization problem in two different ways: The deterministic approach: We formulate an optimization problem that would correspond to the patient’s treatment, using an EDOS system as a constraint. The learning approach: We formulate the optimization problem where the constraint, in this case, will be a neural network previously fitted to a pre-selected data collection. Finally, we will solve it using the methods corresponding to each of the proposed approaches.
In this work we are going to dealt with the design of mathematical frameworks useful in the diagnosis, treatment and possible prognosis of diseases. We will focus on a model for prostate cancer. First, we will pose the optimization problem in two different ways: The deterministic approach: We formulate an optimization problem that would correspond to the patient’s treatment, using an EDOS system as a constraint. The learning approach: We formulate the optimization problem where the constraint, in this case, will be a neural network previously fitted to a pre-selected data collection. Finally, we will solve it using the methods corresponding to each of the proposed approaches.