Novel Method for classification of Hepatitis C Using Support Vector Machine Classifier

Abstract

Aim: The aim of this study is to figure out the presence of Novel Hepatitis C Detection using modern classifiers, and comparing the accuracy, sensitivity, specificity between SVM (Support Vector Machine) and K-NN (K-Nearest Neighbour) Classifiers. Materials and Methods: In this study, data was gathered via the kaggle website. According to clinicalc.com, samples were taken into account as $\boldsymbol{(\mathrm{N}=22)}$ for SVM and $\boldsymbol{(\mathrm{N}=22)}$ for K-NN, with the total sample size being determined using the following parameters: enrollment ratio of 0.1, 95% confidence interval, G power of 80%, and alpha error-threshold value of 0.05. With a standard data set, the accuracy, sensitivity, and specificity were calculated using MATLAB. Results: Independent sample t test SPSS software compares accuracy, sensitivity, and specificity. Between the K-Nearest Neighbor Classifier and Support Vector Machine Classifier, there is a statistically significant difference. In comparison to SVM, the K-NN performed better with $\boldsymbol{\mathrm{p}=0.026}$, p<0.05 accuracy (0.42%), $\mathbf{p=0.021}$, p<0.05 sensitivity (0.43%), and $\boldsymbol{\mathrm{p}=0.001, \mathrm{p} < 0.05}$ specificity (0.43%). Conclusion: K-NN showed better accuracy, sensitivity, specificity than SVM to predict Novel Hepatitis C Detection in a faster way.