Procesamiento de lecturas de actividad motriz para el desarrollo de un modelo de clasificación de pacientes con depresión y personas sanas.

Depression is a mental disorder that can become chronic and significantly hamper the performance of daily life. In its most serious form, it can lead to suicide. Motor activity measurements have become an emerging topic in the field of mental health. Several studies use sensors to measure movements...

Descripción completa

Detalles Bibliográficos
Autores: Espino Salinas, Carlos Humberto, Gamaliel Moreno Chávez, Enrique García Ceja
Tipo de recurso: tesis de maestría
Estado:Versión publicada
Fecha de publicación:2021
País:México
Institución:Universidad Autónoma de Zacatecas
Repositorio:Repositorio Institucional Caxcán
Idioma:español
OAI Identifier:oai:http://ricaxcan.uaz.edu.mx:20.500.11845/2587
Acceso en línea:http://ricaxcan.uaz.edu.mx/jspui/handle/20.500.11845/2587
Access Level:acceso abierto
Palabra clave:INGENIERIA Y TECNOLOGIA [7]
Depresión
Diagnostico
Actividad motriz
Redes neuronales Profundas
Aprendizaje automático
Depresión dataset
Descripción
Sumario:Depression is a mental disorder that can become chronic and significantly hamper the performance of daily life. In its most serious form, it can lead to suicide. Motor activity measurements have become an emerging topic in the field of mental health. Several studies use sensors to measure movements of patients over time to create a diagnosis. The aim is to develop a model based on some machine learning techniques and genetic algorithms, to classify patients with depression and healthy people using motor activity. Readings of 55 patients (32 control patients and 23 patients with condition) were selected, during period of one week, obtaining a total of 385 observations (participants) and 1440 characteristics (time intervals) from which the intervals of one minute more representative to develop a machine learning model using algorithms such as: neural networks, logistic regression, random forests, vector support machines and deep neural networks where the latter obtained the best performance with 80.24% precision, it was trained with 270 of participants and was tested with remaining 30% of data, which 61 were correctly diagnosed as healthy and 32 with depression. Based on these results, it can be concluded that the implementation of these models in an assisted diagnostic tool can help avoid cases of severe depression.