Answers are grounded in this paper and cite their sources.
Data mining techniques predict students’ performance by identifying significant features, suggesting improvement methods.
Educational data mining is a research field that probes undercover patterns in educational data. In this paper, machine learning algorithms have been applied to the dataset that consists of major features so as to predict students’ final grade performances. Thus, the most significant features and the highest-performance machine learning algorithm have been also tried to be detected. To this end, univariate feature selection, tree-based feature selection, and L1-based feature selection methods have been used for the feature selection process. Classification and regression trees, k-nearest neighbors, naive Bayes, random forest, and support vector machines have been employed to build the learning models. The L1-based feature selection and classification and regression trees have delivered the best performance for the feature selection and the model creation processes, respectively. The experimental results demonstrate that the proposed model reached a classification accuracy of 0.7700 and an F1-score of 0.7888 on average. The L1-based feature selection method has selected only 4 features: these are scholarship type, total salary, transportation to the university, and cumulative grade point average in the last semester. In consequence, there exist lots of indicators that impact students' academic successes, the success or failure that emerges after the measurement process can be estimated by regarding these features in advance. Such a task will enable the relationship mechanism between the educational inputs and outputs to be understandable and eliminate shortcomings concerning the education process.
No takes yet. Share an insight, caveat, or question.
Furkan Aydın (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: