Predicting academic performance of university students from multi-sources data in blended learning

In this paper, we propose to predict academic performance of university students from multi-sources data in multimodal and blended learning environments using data fusion and data mining. We have gathered data from 65 university students and different variables from four different sources. Firstly, we apply data fusion and preprocessing for creating a summary dataset in numerical and categorical format. Then, we have applied different white box classification algorithms provided by Weka data mining tool in order to select the best algorithm. Finally, we show the best predicting model in order to help instructor to take remedial actions with students at risk of dropout or failing.