Data Mining Technology in the International Student Evaluation Using Artificial Neural Network with Latent Dirichlet Allocation

Recently, Educational Data Mining (EDM) has increased remarkable consideration. The significant aim of educational institutions is to offer efficient education for students to improve academic performance. When data mining approaches are used in the context of education, they can discover hidden knowledge and patterns that can aid in decision-making processes focused at improving an educational system. Forecasting the student's achievements is a significant problem for higher education administrators. To address this challenge, this research proposes the Machine Learning (ML) method of Artificial Neural Network (ANN) with the Latent Dirichlet Allocation (LDA) approach for the prediction of the performance of the international student's academics. The presented approach considered the last semester for the international student's performance as well as their personal features were examined. The pre-processing techniques like cleaning, integration, transformation, feature scaling as well as reduction were utilized for the prevention of anomalies as well as high dimensionality reduction in the collected data and to acquire the most correlated features. The experimental results shows that the proposed ANN-LDA approach achieves the accuracy of 95.7% when compared to the previous works like K-Nearest Neighbor (KNN) and Support Vector Machine (SVM).