Anomaly detection in ECG using recurrent networks optimized by modified metaheuristic algorithm

Abstract

Cardiovascular disorders, a leading cause of death, demand urgent research attention. Early detection systems hold the potential to improve patient outcomes by enabling timely interventions and lifestyle adjustments. Recent advancements in artificial intelligence algorithms show promise in addressing complex challenges. This study investigates the application of Recurrent Neural Networks (RNNs) optimized with metaheuristic algorithms for anomaly detection in electrocardiogram (ECG) signals. We conducted a comparative analysis of state-of-the-art metaheuristic algorithms to determine their effectiveness in selecting optimal hyperparameters for RNN models, achieving acceptable accuracy levels. Notably, the relatively new crayfish optimization algorithm (COA) is included in the comparative analysis and has exhibited the best overall performance, demonstrating its potential for enhancing cardiovascular disorder detection.