Accurate detection of cardiovascular diseases requires automatic ECG signal monitoring. Semisupervised method which is a combination of automatic monitoring and analysis by an expert cardiologist is considered for the prevention of error in detection. Ventricular arrhythmias were detected to classify ventricular beats earlier. An extension to this method is obtained by classifying the atrial beats. Atrial arrhythmias like Atrial Fibrillation and other abnormalities like First degree block, Atrial Tachycardia, Ventricular Tachycardia is obtained. A Switching Kalman Filter(SKF) is used and a third mode known as X-factor which consists of unknown characteristics is also considered for quality detection. MIT-BIH arrhythmia database is used. The F1 score of 98.11% and Ja of 98.08% were obtained.