الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 116 |  |  |
| | | from | 116 | | |
Abstract
The academic performance of students is an essential factor that influences the accomplishment of any educational institution. Nowadays, predicting students’ academic performance at an early stage of a semester is one of the most crucial research topics in the field of Educational Data Mining (EDM).High Failure rates and dropouts in some courses among students through undergraduate programs are common problems that affect the educational institution’s reputation. So, EDM is used to analyze students’ data gathered from various educational settings to predict students’ performance for taking early actions to enhance the performance of students and also to reduce the failure rates at the end of a semester. The main goal of this thesis is to explore the efficiency of Deep Learning in the field of EDM especially in predicting students’ academic performance to identify the students at risk of failure. A dataset collected from a public 4-year university is used in this thesis to develop predictive models to predict students’ academic performance of upcoming courses given their grades in the previous courses of the first academic year using Deep Neural Network (DNN) and some traditional machine learning algorithms such as Decision Tree, Random Forest, Gradient Boosting, Logistic Regression, Support Vector Classifier, and K-Nearest Neighbor. We have used two model validation methods, which are random hold-out and stratified 5-fold cross-validation. Also, we made a comparison between various resampling methods to solve the imbalanced dataset problem such as SMOTE, ADASYN, ROS, and SMOTE-ENN. from the results, it is observed that the proposed DNN model can predict students’ success in a “Data Structure” course and also can identify the students at risk of failure at an early stage of a semester with an accuracy of 89%.