Cardiac arrhythmias are common heart rhythm disorders that are not always detected during short clinical examinations. Since many arrhythmic events occur intermittently, prolonged cardiac monitoring can improve the identification of abnormal rhythm patterns requiring further evaluation. Single-lead ECG recordings represent a practical solution for long-term monitoring, as they are easy to acquire while retaining essential information for rhythm analysis, including beat timing and fundamental QRS morphology. This thesis presents an interpretable rule-based algorithm for the detection of arrhythmic patterns from single-lead ECG signals. The proposed method integrates features derived from heart rhythm dynamics and QRS complex morphology, with the aim of supporting screening and early detection rather than providing a definitive clinical diagnosis.