Résumé:
In this project, we address the medical field related to the cardiovascular system and
heart diseases. Our objective is to develop a solution based on artificial intelligence
techniques, in particular those of Machine Learning, in the form of an intelligent diagnostic
support system to detect heartbeat anomalies using the dataset of the 2016 PhysioNet/CinC
Challenge. For this we apply techniques for extracting temporal characteristics from PCG
signals. In order to train the model to recognize heart sounds, we used various algorithms such
as Random Forest, KNN and SVM. The results obtained are very satisfactory with an
accuracy score of 92% and demonstrate the effectiveness and reliability of our intelligent
system, which aims to be a diagnostic aid tool for health practitioners.
