Abstract
It is of interest to discuss the feasibility for data mining technology enabled antiretroviral therapy (ART) for HIV positive patients at the University of Gondar Specialized Teaching Hospital, Ethiopia. The Knowledge Discovery in Databases (KDD), which is in an iterative process where evaluation measures enhanced, is used to prepare the data set for ART in HIV positive patients. A decision tree J48 model that is the implementation of algorithm ID3 (Iterative Dichotomiser 3) developed by the WEKA project team is used in this study. The J48 model was built with pruned and without pruned parameters by selecting two different test modes for 10-fold cross validation with percentage split. Results using J48 pruned decision tree with 10-fold cross validation produces 80.5% prediction precision for ART starter prognosis enabled treatment in clinical settings.