Comprehensive Evaluations of Student Performance Estimation via Machine Learning

Success in student learning is the primary aim of educational system. Artificial intelligence utilizes data and machine to achieve excellence learning. In this paper, we exploit several techniques estimate early performance. Two main simulations are used for evaluation. The first simulation Traditional Machine Learning Classifiers (TMLCs) applied House dataset, they Gaussian Naïve Bayes (GNB), Support Vector (SVM), Decision Tree (DT), Multi-Layer Perceptron (MLP), Random Forest (RF), Linear Discriminant Analysis (LDA), Quadratic (QDA). best results were achieved with MLP classifier a division 80% training 20% testing, an accuracy 88.89%. fusion these seven classifiers was also highest result equal MLP. Moreover, second simulation, Convolutional Neural Network (CNN) utilized evaluated on five datasets, namely, House, Western Ontario University (WOU), Experience Application Programming Interface (XAPI), California-Irvine (UCI), Analytics Vidhya (AV). UCI dataset subdivided into three UCI-Math, UCI-Por, UCI-Fused. AV has targets which Math, Reading, Writing. at 97.5%, 99.55%, 98.57%, 99.28%, 99.40%, 99.67%, 92.93%, 96.99%, 96.84% WOU, XAPI, UCI-Fused, AV-Math, AV-Reading, AV-Writing respectively, under same protocol system demonstrates that proposed CNN-based method surpasses all conventional methods other state-of-the-art-work.