Extracting Knowledge in Large Synthetic Datasets Using Educational Data Mining and Machine Learning Models

Abstract

Educational Data Mining (EDM) and Learning Analytics (LA) investigation has emerged as an attractive domain of study. The valuable unfolding experience from institutional databases for several determinations such as prophesying learners achievement rate, enforcement, coordination and improving the teaching–learning manner. The principal intention of learning organizations is to impart high-quality knowledge to their students. One way to produce quality instruction in the education arrangement is by identifying knowledge within EDM. It is the method to estimate the student’s educational achievement. EDM can be explained as the judgment, acquisition, interpretation and broadcasting of data about apprentices and their circumstances, for the persistence of conclusion and optimizing knowledge. Several researchers explored the prediction system with the help of data mining according to generated rules and policies. This research aims to propose a methodology for modelling and implementing algorithms over data from student performance evaluation using machine learning algorithms. This system first collects the data of various students in the entire academic year then applies various feature extraction and feature selection strategies with multiple machine learning algorithms. The experimental analysis has done with Weka environment using a student performance dataset that obtained good classification accuracy for various machine learning algorithms.KeywordsEducational data miningFeature extractionFeature selectionSupervised learningClassificationWeb mining