A Framework to Predict the Student’s Performance in Programing Courses
Download
Author:
WAQAR-UN-NISA
Citable URI :
https://vspace.vu.edu.pk/detail.aspx?id=343
Publisher :
Virtual University
Date Issued:
7/4/2020 12:00:00 AM
Abstract
Academic grades prediction is considered as one of the hot research areas since last decade,
which comes under the domain of educational data mining. It has been observed that in
undergraduate computer science programs, programming courses are considered challenging.
This results in higher tendency of earning lower grades, failures or drop-outs than other computer
science subjects. An early prediction of the students who have high probability of failure (known
as at-risk students) will enable the instructors to intervene and provide extra guidance to learners.
An accurate prediction of student’s grades can directly influence the overall quality of any degree
program and the retention rate of the institution. This research presents a machine learning based
classification model for undergraduate students grades prediction, enrolled in any programming
course(s) in traditional education system. The proposed model is built after careful collection and
pre-processing of data, appropriate feature selection, and model evaluation based on four metrics
namely accuracy, precision, recall and F1-score. Six widely used supervised machine learning
techniques including Random Forest, Artificial Neural Network, K-Nearest Neighbors, Naïve
Bayes, Ordinal Regression, and Support Vector Machine are used after tuning and optimization.
The data used for this research is collected from a private sector university in Lahore. The
collected data covers two major domains: student’s academic record and demographic data. The
results show that Support Vector Machine and K-Nearest Neighbors give highest scores (ranging
from 81% to 94%) for all the evaluation metrics and for all the seven programming courses
considered for this study.
URI :
https://vspace.vu.edu.pk/details.aspx?id=343
Citation:
Nisa,WU(2019).A Framework to Predict the Student’s Performance in Programing Courses. Virtual University of Pakistan(Lahore, Pakistan)
Version :
Final Version
Terms of Use :
Detailed Terms :
Journal :
Files in this item |
Name |
Size |
Format |
Fall 2019_CS720_MS150400120.pdf |
1555kb |
pdf |