Analysis of Associative Classification for Prediction of HCV Response to Treatment

Abstract

objective of this research is the analysis of predicting the response for treatment in patient with hepatitis C virus. The Interferon Alfa (IFN) in combination with ribavirin (RBV) is used as a standard therapy for chronic hepatitis C (CHC), it is very expensive and accompanied with great side effects, with that it fails in more than half cases. For the prediction of treatment response, a knowledge discovery framework including two main phases: pre- processing and data mining was developed. In pre- processing phase, the cleaning and selection of suitable features from patients' data were done. In data mining phase the selected patients' features were mined using Associative Classification (AC) technique to generate a set of Class Association Rules (CARs). The most suitable rules from the generated CARs were selected to build a classifier, which predicts patients' response for treatment. Using our classification model, 220 patients treated with IFN plus RBV were analyzed, 92 patients resulted responders and 128 non- responders at the end of treatment and during the follow up. 170 cases had been used to train our intelligent models and 50 patients had been used to test the models. The experiment results showed that the proposed technique is an effective classification technique with high prediction accuracy reach up to 90%.