An support vector machine model for classification of premature cardiac contractions

It is difficult to determinate the premature cardiac contraction type of a case in clinic due to its vague signals in electrocardiogram.As an approach of computer-assisted diagnosis,a model for classification was proposed based on support vector machine(SVM).All samples data were derived from 82 clinic cases.By means of our SVM model,the accuracies of classification were up to 94.44%for the training set and 92.86%for the testing set.The accuracy of leave-one-out cross-validation was 92.59%. The satisfactory results indicate that the proposed approach is effective and could be applied to assisted diagnosis in clinic practice.