Prediction of ECG Cardiac Abnormality Signal using Supervised Prediction Model

The detection of a patient's cardiac abnormal activity is crucial when an abnormal symptom occurs. The objective of this initiative is to detect disease using some supervised prediction model. Several data are extracted from the electrocardiogram (ECG) signal and used as input parameters which are amplitude and duration of P peak, amplitude, and duration of QRS wave and amplitude and duration of T peak. The proposed prediction models are K-Nearest Neighbours (KNN), Discriminant Analysis, Principal Component Analysis (PCA) and Decision Tree. The obtained result is then compared to other prediction models to identify the optimal performance based on the accuracy prediction and lowest mean square error (MSE). It shows that KNN prediction model outperforms other models with 94.27% accuracy and 0.24 on the MSE measurement.