Algorithm 1 Pseudocode of the PSO-based feature selection procedure used to determine the optimal subset of discriminative ECG features.

Input: Feature matrix (X), Class labels (y) Output: Optimal feature subset (gBest) Procedure PSO_Feature_Selection()     Initialize the particle population with random binary feature subsets     Initialize particle velocities and PSO parameters (ω, c1, c2)     while termination condition not met do         for each particle in the population do             Evaluate fitness using selected features (e.g., classification accuracy)             Update personal best position (pBest) and global best (gBest)             Update velocity and position using standard PSO equations             Apply sigmoid transfer function to binarize new positions         end for         Update global best solution if needed     end while Return: gBest → selected optimal feature subset