|
Document Type |
: |
Thesis |
|
Document Title |
: |
USING DATA MINING TECHNIQUES TO PREDICT STUDENTS ACADEMIC PERFORMANCE استخدام تقنيات تنقيب البيانات للتنبؤ بالأداء الاكاديمي للطلاب |
|
Subject |
: |
Faculty of Computing and Information Technology |
|
Document Language |
: |
Arabic |
|
Abstract |
: |
Ability to predict students’ academic performance is critical for any educational institution aims to improve their students' performance and perseverance. Although students’ performance prediction studied widely, it still represents a challenge and complex issue because student performance influenced by different features. Previous works have studied the impact of different features on students' performance, but few works have focused on the impact of students’ assessment grades and online activity on Learning Management System (LMS) together.
This research aims to investigate the impact of assessment grades and online activity data jointly and separately to determine their effect on students’ performance. By employing sub-datasets to create prediction models for students performance. Based on five classification algorithms are decision tree, random forest, sequential minimal optimization, multilayer perceptron, and logistic regression. In addition to feature selection task based on filter and wrapper methods for identifying the most important features that affect student performance. This work carried out using data extracted from LMS for 241 undergraduate students in Information Systems department, Faculty of Computing and Information Technology at King Abdulaziz University.
Results showed that assessment grades is the important features that affect students’ performance, especially assignments mark and final exam. Moreover, prediction models created based on assessment grades separately or jointly with activity data were perform better than models using activity data alone. That indicate assessment grades had a greater impact on student performance, while activity data had a lower impact. However, models that include assessment and activity data together work well, which means it is important to include assessment grades with activity data for predicting students’ performance. Random forest performed better than other algorithms, followed by a decision tree. |
|
Supervisor |
: |
Dr. Bassam Zafar and Ahmed Mueen |
|
Thesis Type |
: |
Master Thesis |
|
Publishing Year |
: |
1441 AH
2020 AD |
|
Added Date |
: |
Sunday, July 12, 2020 |
|
Researchers
| أمل مهدي ال حسان | Alhassan, Amal Mahdi | Researcher | Master | |
|