Abstract
IL28 single nucleotide polymorphism (rs12979860) is an aetiology-independent predictor of hepatitis C virus (HCV) related hepatic fibrosis. Data mining is a method of predictive analysis which can explore tremendous volumes of information found in health records to discover hidden patterns and relationships. The study aims to evaluate and compare the prediction accuracy of APRI, FIB-4 versus data mining for prediction of HCV related advanced fibrosis. This retrospective study included 427 patients with chronic HCV. We used the data mining analysis to construct Reduced Error Pruning (REP) decision tree then applied Auto-WEKA to select the best classifier out of 39 algorithms to predict advanced fibrosis. FIB-4 and APRI had sensitivity and specificity of 0.415, 0917, 0.523 and 0.831 respectively. REP-tree algorithm was able to predict advanced fibrosis with sensitivity 0.749, specificity 0.729 and ROC area 0.796. Out of 16 attributes, IL28 genotype was selected by the REP-tree as the best predictor of advanced fibrosis. Using Auto-WEKA, the Multilayer perceptron (MLP) neural model was selected as the best predictive algorithm with sensitivity 0.825, specificity 0.811 and ROC area 0.880. Thus, MLP is better than FIB-4, APRI and REP-tree in predicting advanced fibrosis in patients with chronic HCV.
