Guilherme Cunha Santos, AntonioAntonioGuilherme Cunha SantosLobo Marques, Joao AlexandreJoao AlexandreLobo MarquesRigo Jr., LuisLuisRigo Jr.Paulo Madeiro, JoaoJoaoPaulo Madeiro31/12/202331/12/2023202320232023https://dspace.usj.edu.mo/handle/123456789/542610.22489/CinC.2023.312In the George B. Moody PhysioNet Challenge 2023 on ‘Predicting Neurological Recovery from Coma After Cardiac Arrest’, our team, UF_MDCC, employed machine learning techniques to predict patient prognosis based on electroencefalogram (EEG) signals. Our strategy was to extract features from the EEG signals, capturing both linear and non-linear characteristics from time and frequency domains. The chosen model was Random Forest, trained with various feature extraction strategies. Our team's performance on the test set at different time intervals are as follows: At 12 hours - Rank 10, Challenge Score 0.312; at 24 hours - Rank 24, Challenge Score 0.312; at 48 hours - Rank 28, Challenge Score 0.272; and at 72 hours - Rank 32, Challenge Score 0.272.EnglishEEG Signal Featurestext::conference output::conference proceedings::conference paper