Measuring Students’ Performance with Data Mining

Abstract

Understanding the true reasons behind students’ failure, and bringing preventive measures to this issue at early stages are invaluable in the educational learning process. Preventing problems such as language deficiency or misclassification of the students in the appropriate academic levels is primordial for any educational institution. Many factors influence the learning process of the students, such as the demographic characteristics, educational background as well as language barrier. This work highlights the most preponderant factors affecting students’ advancement in the learning process and provides support to academic administrators. It uses some of state of the art classification and regression algorithms in the application domain of predicting students’ progress. Datasets were filtered and trained using predictive algorithms. It is shown that Science learning and English language skills are highly correlated. Datasets are not always suitable for data mining unless it is preprocessed and well adapted to the context being studied. A tool has been developed to preprocess the data provided that feeds into Weka Data Mining Software to profile students’ performance.