Inteligencia Artificial Explicable para estimar la depresión y esquizofrenia en pacientes basadas en datos de sensores de IoT
Loading...
Official URL
Full text at PDC
Publication date
2023
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citation
Abstract
Actualmente, la Inteligencia Artificial puede lograr distintas tareas como la toma de decisiones y la resolución de problemas. La Inteligencia Artificial Explicable (XAI) es un conjunto de procesos y métodos que permite a los humanos confiar y entender la salida de los algoritmos de aprendizaje automático.
En el presente trabajo se realizan distintos métodos de clasificación y predicción a partir de dos conjuntos de datos que fueron recogidos por sensores incorporados en un reloj actigráfico en pacientes con trastorno depresivo o esquizofrénico. Posteriormente, estos modelos son explicados mediante las dos librerías de XAI m ́as conocidas, SHAP y LIME. Además, se realiza una comparación con otros artículos escritos a partir de estos mismos datasets.
Nowadays, Artificial Intelligence can accomplish different tasks such as making decisions and solving problems. Explainable Artificial Intelligence (XAI) is a set of processes and methods that allow humans to trust and understand the output of machine learning algorithms. In the work, different classification and prediction methods are performed on two datasets that were collected by sensors incorporated in an actigraphic clock in patients with depressive or schizophrenic disorder. Subsequently, these models are explained using the two most well-known XAI libraries, SHAP and LIME. In addition, a comparison is made with other articles written from these same datasets.
Nowadays, Artificial Intelligence can accomplish different tasks such as making decisions and solving problems. Explainable Artificial Intelligence (XAI) is a set of processes and methods that allow humans to trust and understand the output of machine learning algorithms. In the work, different classification and prediction methods are performed on two datasets that were collected by sensors incorporated in an actigraphic clock in patients with depressive or schizophrenic disorder. Subsequently, these models are explained using the two most well-known XAI libraries, SHAP and LIME. In addition, a comparison is made with other articles written from these same datasets.
Description
Trabajo de Fin de Máster en Internet de las Cosas, Facultad de Informática UCM, Departamento de Ingeniería de Software e Inteligencia Artificial, Curso 2022/2023