Early Prediction of students’ grades and appropriate Recommendations
DescriptionEducational data mining techniques are widely used in academic prediction on student performance in classroom education. In this research, we performed an analysis to identify the significant impact of student background, student social activities and student coursework achievement in predicting student academic performance. By using the combination with all the attributes and comparing it with different classification tools Decision Tree, Random Forest, and SVM, student coursework achievement is the most valuable attribute to predict the student’s performance. Moreover, SVM can have the highest accuracy as the prediction model(0.913). Here, we also proposed a recommendation system that recommends courses for students based on similarities of students’ course history. The proposed system employs data mining techniques to identify patterns in the course enrollment procedure. We also have noticed that clustering the students based on their history of enrollment made the association rules generated using apriori most efficient than the rules generated without clustering. The frequent items that were found from the rules are depicted as a network. The efficiency of this network is verified by distributing its degree. The result of this is a scale-free network.