Publication: Seguimiento automatizado de variables de control en pacientes crónicos
Loading...
Official URL
Full text at PDC
Publication Date
2016
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
Un sueño escaso o de mala calidad puede tener repercusiones negativas en la vida diaria. Si esta situación se prolonga en el tiempo, puede afectar al estado de ánimo e interferir en el trabajo y la vida social de una persona. Son muchos los factores que determinan la calidad del sueño. No es suficiente con dormir, sino que hay que dormir bien para que el sueño sea realmente reparador y se pueda uno levantar descansado y con energía. Muchos aspectos como la actividad física, las comidas y bebidas, los horarios, pueden alterar la calidad del sueño o generar dificultades para conciliarlo por lo que en los últimos años han salido al mercado varios dispositivos y aplicaciones cuantificadoras que ayudan a monitorizar el sueño. Existen dispositivos que solamente registran el movimiento y la temperatura corporal y aun así generan una gran cantidad de datos. Debido a esto, la información producida puede no ser fácilmente interpretable. En este proyecto usamos un dispositivo de monitorización que ofrece más información (como temperatura y flujo térmico) por lo que queremos mejorar la interpretabilidad de los datos que se generan para ayudar a facilitar el trabajo de los médicos que las utilizan con sus pacientes intentando descubrir trastornos metabólicos. Se aspira a simplificar el trabajo manual que ahora se realiza intentando que sea más automático.
Para ello, en este proyecto se utilizan técnicas de minería de datos en series temporales que permiten hacer un análisis exploratorio de los datos y agrupar la información de forma que los resultados se muestren de una forma resumida y concreta y que pueda ser percibida de un vistazo. Utilizando estas técnicas creamos una aplicación que obtiene series temporales de los datos particionadas en distintos tipos de episodios, centrándonos en episodios de sueño y de actividad física, clasificando los episodios de sueño por su similitud y presentando un resumen de la actividad del paciente a lo largo de toda la monitorización.
Sparce or poor quality sleep can have a negative impact on daily life. If this situation continues over time, it can affect the mood and interfere with work and social life. There are many factors that determine the quality of sleep. It is not sufficient with sleeping every day, but must sleep well so that sleep is a really restful sleep and one can lift rested and energized. Many aspects such as physical activity, meals and drinks, schedules, can disrupt sleep quality or cause difficulty falling asleep so in recent years have come on the market several bracelets or armbands and quantifying applications that help monitor the sleep. There are devices that only record the movement and body temperature and yet generate a large amount of data. Because of this, the information produced can not be easily interpretable. In this project we use a monitoring device that provides more information (such as temperature and heat flow) so we want to improve the interpretability of the data generated to help ease the work of doctors who use it with their patients trying to discover metabolic disorders. We aim to simplify the manual work that is now performed trying to be more automated. For this purpose, in this project time series data mining techniques are used that allow an exploratory analysis of the data and group information so the results are displayed in a summarized and concrete way and allow to observe them at a glance. Using these techniques, we have created an application that obtains time series data partitioned into different types of episodes, focusing on episodes of sleep and physical activity, classifying sleep episodes by their similarity and a doing a summary of patient activity throughout of all monitoring.
Sparce or poor quality sleep can have a negative impact on daily life. If this situation continues over time, it can affect the mood and interfere with work and social life. There are many factors that determine the quality of sleep. It is not sufficient with sleeping every day, but must sleep well so that sleep is a really restful sleep and one can lift rested and energized. Many aspects such as physical activity, meals and drinks, schedules, can disrupt sleep quality or cause difficulty falling asleep so in recent years have come on the market several bracelets or armbands and quantifying applications that help monitor the sleep. There are devices that only record the movement and body temperature and yet generate a large amount of data. Because of this, the information produced can not be easily interpretable. In this project we use a monitoring device that provides more information (such as temperature and heat flow) so we want to improve the interpretability of the data generated to help ease the work of doctors who use it with their patients trying to discover metabolic disorders. We aim to simplify the manual work that is now performed trying to be more automated. For this purpose, in this project time series data mining techniques are used that allow an exploratory analysis of the data and group information so the results are displayed in a summarized and concrete way and allow to observe them at a glance. Using these techniques, we have created an application that obtains time series data partitioned into different types of episodes, focusing on episodes of sleep and physical activity, classifying sleep episodes by their similarity and a doing a summary of patient activity throughout of all monitoring.
Description
Trabajo de Fin de Grado en Ingeniería Informática (Universidad Complutense, Facultad de Informática, curso 2015/2016)