Analysis of Various Machine Learning Algorithm for Cardiac Pulse Prediction

By applying machine learning algorithms, patterns are identified or recognized in the process of Pattern Recognition. On the grounds of prior knowledge, the data is collected and sorted. In this method, the raw data is transformed into a susceptible form which can be used by the machine. Electrocardiogram (ECG) Pattern Recognition is the main focus of this paper. ECG keeps a track of heart's electrical activity. In the field of biometric it is used as a robust biometric. On the person, off the person and in the person, are the three categories for tracking and capturing signals. Only Off-the-person category in which there is no or minimal skin contact, is included in this paper. To analyze and implement data, six baseline methods are utilized. These baseline methods are applied two publicly available databases-CYBHi and UofT. Raw signals and spectrogram of heartbeat are used for studying about representing features. Various machine learning algorithms are also discussed. Implementation for predicting heartbeat as normal or abnormal and heart diseases, is performed.