ABSTRACT Name : Nayana Taradhanti Hodi Program : Computer Science Title : Predicting Student’s Risk of Failure based on Transcript using Ensemble of Classifiers Ensemble classification is a method that combine classification’s models to achieve greater predictive accuracy than single model. This study aims to get risk prediction model with high accuracy using ensemble combination method. Risk prediction model produced based on transcript contains student’s academic history data. This data, then processed to get the wanted attributes. Ensemble combination method that is use to produced risk prediction model is majority voting and stacked generalization. Testing is done twice using different training and testing data. The produced model is then compares with single model, to gain model with better accuracy. On the first test, stacked generalization is a model that performs best and stable with different datasets. On the second test, ensemble classification test is not the best model, but also not the worst. It can be said, ensemble classification method helps single method with lowest accuracy to perform better. Keywords: risk prediction, ensemble, ensemble classification, learning analytics