Predictive Modeling Algorithms-based Classification of Arrhythmia

Arrhythmia, change in rate or rhythm of the heart. Classification and Prediction of arrhythmia accurately is a tedious work in the present scenario. The proposed work aims to classify normal and abnormal heart beat by identifying the relationship between the output and input variables. It also predicts the type of arrhythmia such as bradycardia, tachycardia and normal sinus rhythm. A Hybrid combination of Supervised Machine Learning Algorithms and Classification techniques is employed in classification and prediction of arrhythmia. The proposed work compares the accuracy of six different supervised learning algorithms and the results from best algorithm is given as input to find the type of Arrhythmia. In terms of accuracy Logistic regression achieves an accuracy of 86.21% in the test dataset compared to all other algorithms. After results from classification, Peaks of the waves are detected and type of arrhythmia is found. It is found that 43% of persons are normal and 57% of persons are affected with arrhythmia.