Novel and Efficient Classification of Cardiovascular Abnormalities by Machine Learning

Abnormality detection of ECG signal is useful for classifying cardiovascular problems. The most popular techniques for identifying abnormalities in ECG signal is Arrhythmic beat classification. Wavelet transform and Principal component analysis (PCA) and Wavelet transform are applied to ECG signal to extract morphological, spectral features and wavelet features. ECG signal processing and Machine learning classifier based arrhythmic beat classification are implemented to classify into abnormal and normal subjects in proposed research. Discrete wavelet transform and PCA are used to extract feature points in ECG signal and we used Random Forest (RF) classifier and Multinomial Logistic Regression (MLR) classifier for training and testing by <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{5}$</tex>-fold cross validation to asset performance and to classify abnormality of ECG signal obtained from patient heart. This proposed work is tested on MIT-BIH arrhythmia public database and overall Accuracy of multinomial logistic regression classifier is (99.6 %) high compared to RF classifier but true positive rate for <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{R F}$</tex> is higher than Multinomial logistic regression. We also conducted experiment and our methodology showed improved results while compared with other machine learning algorithms

Перевод пока недоступен