Applying Supervised Machine Learning Algorithms to Detect Cardiac Events

Within the realm of machine learning, numerous research advancements have enhanced the understanding of data analytics and prediction models. One of the more recent achievements in artificial intelligence is the rise of machine learning in healthcare, aiding in the development of streamlined treatment and diagnosis. Cardiac focus in this paper is due to an interest in how the pandemic restricted extracurriculars and athletics in school, which led to a decrease in physical activity in adolescents. With a decrease in physical activity, the cardiac systems of students might have weakened thus fostering an interest in applying machine learning to cardiac health in adolescents. By using wearable devices and mobile devices to collect data from participants (mainly adolescents), machine learning algorithms can be applied to the data and then analyzed to get information about the cardiac states of adolescents. Cardiac features were measured using the YAMAY Smart Watch wearable device; a variety of supervised machine learning algorithms (KNN, Naïve Bayes, Random Forest, and Decision Trees) were used to predict the expected data with the target data. Overall, after testing each of the supervised machine learning algorithms, Random Forest had the best prediction accuracy of 75.86%. With these results in mind, research focusing on applying supervised machine learning algorithms to detect cardiac events would benefit from using Random Forest.