CARDIAC AND STRESS ASSESSMENT USING SVM CLASSIFIER

Electrocardiogram (ECG) is an electrical recording of the heart and used to measure the rate and regularity of the heartbeats. The obtained ECG signals are noisy, due to the loss of signal values, problems in the electrodes and natural addition of noises in the image. Initial filtering of the noise is done using wavelet transform. Features extracted are Mean, Median Moment, Skewness, Standard Deviation, Variance, Frequency, Entropy, Mode, Maximum, Minimum, Phase, Magnitude and Kurtosis. The Rpeaks are detected from the signals and the values are combined along with the previously calculated features. Optimizations are performed with the extracted feature that reduces the number of feature values by selecting the best feature values from the extracted feature values. Then classification is done using Support Vector Machine classifier. The classifier analyzes the feature values and identifies whether the input signal is normal signal or abnormal signal. Keywords: Electrocardiogram (ECG), Particle Swarm Optimization (PSO), Support Vector Machine (SVM), Wavelet Transform (WT).