|
![](images/pix01.gif) |
|
|
|
Integration between Deep Neural Network and Predictive Learning Analytics (PLA): to Improve Student’s Performance in Online Exam |
|
PP: 273-287 |
|
doi:10.18576/isl/130203
|
|
Author(s) |
|
H. F. Balat,
G. K. Hebesh,
Salem Alkhalaf,
M. R. Alkotby,
M. Abdou Amasha,
R. K. Arafa,
|
|
Abstract |
|
In this paper we discussed the Predicting performance of university students has become an important necessity among education experts, as it helps rationalize spending and invest efforts correctly. The current study sheds light on the best techniques for predicting the performance of university students in an online exam conducted over the Internet. The study used the integration between Deep Neural Network and Predictive Learning Analytics (PLA) to improve students performance in these e-exams. The study was applied to a sample of students from the Faculty of Engineering at Damietta University through their scores in the electronic exam during the years from 2019 to 2022. The dataset was divided into train (687 instances (50%)) and exam (686 instances (50%)). Besides, the dataset includes four features, such as year, score, percent, and grade. Five ML algorithms were selected to examine high prediction accuracy. "The ML algorithms are Random Forest (RF), Naive Bayes (NB), Support Vector Machine (SVM), Decision Table (DT), and K-Nearest Neighbors (KNN). Besides, evaluation metrics were applied to compare ML algorithms such as confusion matrix, accuracy, recall, precision, and F-measure. According to the results, the Random Forest, Decision Table, and K-Nearest Neighbors classifiers were the best, as they correctly classified 685 instances during the prediction of the students performance. For other metrices of evaluation (recall, precision, and F-measure), concerning the Random Forest, Decision Table, and K-Nearest Neighbors classifiers, precision, recall, and F-measure achieve 0.99."Our finding also revealed the successful achievement of high accuracy in predicting the performance of students at the Faculty of Engineering, Damietta University, in online exams using ML algorithms. |
|
|
![](images/pix01.gif) |
|
|