Comparative Analysis of Hepatitis C Using K-Nearest Neighbor Classifier and Decision Tree Classifier

Abstract

Aim: The purpose of this study is comparing the accuracy, sensitivity and specificity of K-Nearest Neighbor Classifier and Decision Tree classifier in detecting the presence of Novel Hepatitis C Detection using contemporary methods. Materials and Methods: The kaggle website was used to collect data for this study. According to clinicalc.com, samples were considered as (N=22) for K-NN classifier and (N=22) for decision tree, with an alpha error-threshold value of 0.05, enrollment ratio of 0.1, 95% confidence interval, G power of 80%, and total sample size determined. Using matlab programming and a standard data set, the accuracy, sensitivity and specificity were computed. Results: SPSS software is used to compare accuracy, sensitivity and specificity using Independent sample t-test. Between K-NN and Decision Tree classifiers, there is a statistically significant difference p=0.003,p<0.05-accuracy (0.42%), p=0.003,p<0.05-sensitivity (0.43%), and insignificant difference in p=0.678, p>0.05-specificity (0.43%) in K-NN. The K-NN showed better results in comparison to the Decision Tree. Conclusion: K-NN outperformed Decision Tree in terms of accuracy, sensitivity and specificity in predicting Novel Hepatitis C Detection.