Heart Rate Anomaly Detection in Healthcare Using Elliptic Envelope and Local Forest

Medical professionals use Anomaly Detection(AD) to identify patients with potential health problems in the heart rate data (HRD), an essential metric related to cardiovascular health. In this research, the most effective AD techniques were identified in detecting unusual variations of the heart rate pattern. Synthetic HRD data with anomalous rates of 15%, 20%, and 25% was generated and standardized it using min-max scaling for consistency among the datasets. Normal HR was based on function with a mean of 70 and a standard deviation of 5 while anomalous values were generated with a mean of 120 and a standard deviation of 10. Among the evaluated models (Elliptic Envelope, one-class SVM, Local Outlier Factor, and Isolation Forest), the Elliptic Envelope (EE) and Isolation Forest (IF) models obtain the best scores in terms of recall and precision – 0.92 and 0.89, respectively. Overall, the results of this research indicate that equally robust AD models should be implemented within clinical systems and wearable devices that serve as a digital space for early cardiac risk identification, prompting adequate intervention and, as a result, health improvement.