Predictive Modelling for Student Grades in FYP

• Main Author: Ng, Kerwin
• Format: Final Year Project / Dissertation / Thesis
• Published: 2021
• Subjects: QA76 Computer software
• Online Access: http://eprints.utar.edu.my/4093/1/1600492_FYP_report_%2D_KERWIN_NG.pdf; http://eprints.utar.edu.my/4093/
• Institution: Universiti Tunku Abdul Rahman

Predicting students’ grade in Final Year Project is difficult because the factors may not be purely based on a student’s academic performance. The project focus on using the academic performance of students and their logbook to predict the Final Grades of students in the Final Year Project. This project aims to predict the grade of students in the Final Year Project to decrease the student’s failure, attrition and withdrawal rate. The project proposed using classification which is part of the data mining process to predict the students’ Final Year Project Grades. The proposed prediction model are K-Nearest Neighbours, CART, C4.5, Naïve Bayes, Support Vector Machine and Neural Network. The methodology adopted by the project is a modified version of CRISP-DM (Cross Industry Standard Process for Data Mining) to cater to the needs of this project. The steps include domain understanding, data collection, data understanding, data preparation, modelling and model evaluation.The project successfully created a dataset based on students’ logbook and academic data which will ease future students’ work to do predictions on FYP 2 Grades of students. Empirical studies have been performed and it is found that other than CGPA many features collected during the data collection process are found useful in predicting the Final Grades of students in the Final Year Project. It is also confirmed that the use of Support Vector Machine Model on the dataset created during the project can deliver a good outcome in predicting students FYP2 Grades.