Predicción de la fase pre-ictal de convulsiones en pacientes con epilepsia a partir de señales electroencefalográficas y electrocardiográficas
Seizures are harmful to patients, who, without timely prediction, can lead to death. Therefore, having algorithms that indicate when an epileptic episode is going to occur provides security and action time to act. The present work focuses on the prediction of seizures in patients with epilepsy from...
Guardado en:
| Autor principal: | |
|---|---|
| Otros Autores: | |
| Formato: | Trabajo de grado (Pregrado y/o Especialización) |
| Lenguaje: | spa |
| Publicado: |
Universidad Antonio Nariño
2021
|
| Materias: | |
| Acceso en línea: | http://repositorio.uan.edu.co/handle/123456789/5015 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | Seizures are harmful to patients, who, without timely prediction, can lead to death.
Therefore, having algorithms that indicate when an epileptic episode is going to
occur provides security and action time to act.
The present work focuses on the prediction of seizures in patients with epilepsy
from electroencephalography (EEG) and Electrocardiography (ECG) signals. The
study was carried out in patients who suffered seizures and Machine Learning
algorithms were implemented for the prediction of the pre-ictal phase of seizures
using the "Class Learner" tool from Matlab.
For the development of the work, the CRISP-DM methodology was used, with
which characteristics of 10 patients can be extracted in order to train different
classification algorithms.
The EEG and EKG signals were considered together and separately to show
which of the two obtained better performance according to the metrics computed
from the confusion matrix. It was shown that the best sensitivity was obtained
when the characteristics extracted from the EEG and EKG were worked together. |
|---|
