A large body of literature has been published regarding E-learning and improving students’ performance through the use of specialised advising systems. However, much less research has focussed on the use of data mining for E-learning systems to develop an effective performance prediction and advising tool. This study contributes to the literature in these under-researched areas. Predicting students’ performance in E-learning is not dependable on traditional face-to-face education methods, where the instructor interacts and receives direct feedback from their students. In E-learning systems, student performance prediction is based on analysing students’ marks and progress. This thesis aims to apply data mining to all available data in the E-learning environment so that the process and effectiveness can be improved by predicting students’ performance and suggesting advice to both the instructors and students based on the predictions made. The study thus aims to use data mining in order to improve the accuracy of previous prediction models, by gaining additional insights from student data. Details of previous prediction models and how these have been used or replaced by the new proposed system will also be presented. This study adopts an experimental research approach where the researcher developed and tested an advising e learning software on a sample of students at the University of Dammam. The study has resulted in the production of a software that is able to give students advice based on their unique virtual learning experience, analysing metrics such as number and type of activities undertaken in the learning environment. The advice consists of suggestions regarding the best way to achieve higher predicted results, offering students clear, practical and quantifiable solutions (clearly measurable and having well-set rules and proposed solutions) to achieving higher academic performance. For instance, suggestions can include taking a number of additional quizzes or spending a certain time on a particular exercise. The software was tested on a sample of students and has resulted in the average scores of students significantly increasing by 5% from 67% to 72%. This suggest that the software can be used to improve students’ performance if applied to larger samples
Dyddiad Dyfarnu | 29 Maw 2017 |
---|
Iaith wreiddiol | Saesneg |
---|
Sefydliad Dyfarnu | |
---|
Noddwyr | University of Dammam |
---|
Goruchwyliwr | Richard Jensen (Goruchwylydd) & Qiang Shen (Goruchwylydd) |
---|
Data Mining E-Learning Data for a Student Prediction System
Alghamdi, F. (Awdur). 29 Maw 2017
Traethawd ymchwil myfyriwr: Traethawd Ymchwil Doethurol › Doethur mewn Athroniaeth