Backpropagation algorithm for classification problem: academic performance prediction model for UiTM Melaka Mengubah Destini Anak Bangsa (MDAB) program. / Fadhlina Izzah Saman, Nurulhuda Zainuddin and Khairiyah Md Shahid
Artificial neural networks (ANN) has become one of the artificial intelligent techniques that has many successful examples when applied to classification problem such as doing pattern recognition and prediction. Multilayer perceptrons (MLPs) is one of the topology used for processing ANN, while back...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Research Reports |
| Language: | English |
| Published: |
Research Management Institute (RMI)
2012
|
| Online Access: | http://ir.uitm.edu.my/id/eprint/18216/2/LP_FADHLINA%20IZZAH%20SAMAN%20RMI%2012_5.pdf http://ir.uitm.edu.my/id/eprint/18216/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Artificial neural networks (ANN) has become one of the artificial intelligent techniques that has many successful examples when applied to classification problem such as doing pattern recognition and prediction. Multilayer perceptrons (MLPs) is one of the topology used for processing ANN, while backpropagation algorithm is one of the most popular methods in training MLPs. UiTM Melaka has set one of the Quality Objectives to be achieved for each faculty is to produce at least 65% of full time students graduating with a CGPA of at least 3.00. There is no existing tool to assist faculties in estimating the number of students that can achieve the objective, hence a prediction model using Backpropagation Algorithm is proposed by using a case study of UiTM Bandaraya Melaka Bachelor of Administrative Science students. The initial model will analyze a trend of past students' achievement upon graduation based on factors such as diploma CGPA and 15 core subjects' results, and after a series of experiments, a final model will be obtained with the best parameters to produce the best results. The final model then will produce an output in the form of prediction for current students' graduation CGPA. The output can be used to identify potentially good and weak students, and for the faculty to arrange the teaching and learning session according to students' capabilities in order to produce students with a CGPA of at least 3.00 upon graduation. |
|---|