Abstract
A digital filtering based noise removal, wavelet transform based features extraction and machine learning based classification approach is suggested for automated arrhythmia identification. The system processes the Electrocardiogram (ECG) signals to attain this classification. Features of the filtered signal are minedby using a specific wavelet transformation scheme. In next step, these features are analyzed for the distinguishing of arrhythmia categories. The execution of the framework is analysed by employing a standard database. Results assure that the designed framework is able to effectively transform the ECG signals into features. It renders the handling of lesser information by the post modules. This guarantees an eminent lessening within the data storage, transfer and classification activities. Also, the suggested framework performance in terms of arrhythmia distinguishing is evaluated. Findings are promising and results in 99.4% of classification accuracy.