ECG based Prediction Model for Cardiac-Related Diseases using Machine Learning Techniques
Igor Matias, Nuno Garcia, Miguel Sousa, Eftim Zdravevski
| Abstract: This dissertation presents research on the construction of predictive models for health conditions through the application of Artificial Intelligence methods. The work is thus focused on the prediction, in the short and long term, of Atrial Fibrillation conditions through the analysis of Electrocardiography exams, with the use of several techniques to reduce noise and interference, as well as their representation through spectrograms and their application in Artificial Intelligence models, specifically Deep Learning. The training and testing processes of the models made use of a publicly available database. In its two approaches, predictive algorithms were obtained with an accuracy of 96.73% for a short horizon prediction and 96.52% for long Atrial Fibrillation prediction horizon. The main objectives of this dissertation are thus the study of works already carried out in the area during the last decade, to present a new methodology of prediction of the presented condition, as well as to present and discuss its results, including suggestions for improvement for future development.
In uBibliorum, Repositório Digital da UBI