Heart Rate Arrhythmia Identification with Internet of Things and Machine Learning

Heart rate classification is a critical task in health monitoring and diagnosis, particularly facilitated by advancements in Internet of Things (IoT) and wearable technology. This study evaluates the performance of various machine learning models, with a specific focus on Support Vector Machine (SVM) with a linear kernel, in classifying heart rate data. The testing indicates that SVM with a linear kernel achieves a testing accuracy of 100%, surpassing other models such as Random Forest (99.6%) and Decision Tree (99.3%). This exceptional performance is attributed to the linear separability of the heart rate data, where SVM with a linear kernel effectively identifies the optimal hyperplane for class separation. Additionally, SVM's linear kernel demonstrates robustness against noise, which is common in real-world heart rate data, thereby enhancing its reliability. The study also highlights the interpretability of linear models through feature weights, providing insights into the physiological factors influencing heart rate. This research bridges significant gaps in existing literature by demonstrating the accuracy and practical applicability of linear kernel SVMs in heart rate classification, especially in the context of IoT and wearable technologies. The findings suggest that simpler models should be considered before resorting to complex non-linear models, offering a balance of high performance, computational efficiency, and interpretability in medical applications.