Effect of curve fitting in an automated cardiac arrhythmia detection system

An Automated offline Arrhythmia detection system is proposed with improved accuracy. The raw ECG signal obtained from the existing database is preprocessed using the FFT filter. In order to construct a smooth curve that will reproduce the HRV signal, the degree of the polynomial equation is determined by how many points are needed to match. A polynomial degree of n=7 equation is used which provides the best results in fitting the HRV signal. Wavelet and statistical parameters are combined as a single set of parameters called hybrid with curve fitting. The performance of the suggested algorithm is compared with other algorithms and tested on a MIT database. The accuracy of the proposed offline system is 93.75%.