Sentiment analysis of student's opinion on programming assessment using naive bayes algorithm on small data
Student opinion could be used to facilitate institutions to improve the quality of teaching and learning by delivering the appropriate teaching method based on the student’s learning experience. The purpose of this study is to investigate the efficiency of data mining techniques for the sentiment an...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2019
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/96676/1/MahmodUmarMSC2019.pdf.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Student opinion could be used to facilitate institutions to improve the quality of teaching and learning by delivering the appropriate teaching method based on the student’s learning experience. The purpose of this study is to investigate the efficiency of data mining techniques for the sentiment analysis of student opinion on programming subject assessment. Two machine learning algorithms, which are Support Vector Machine (SVM) and Naïve Bayes (NB) have been identified to be the best in sentiment analysis on large data. SVM performs better than NB on big data but the case may not be the same on small dataset. The research aim is to design a framework that will investigate the efficiency of Naïve Bayes algorithm on two sentiment classification classes namely positive and negative on small dataset. A comparative performance measure is done using SVM and lexicon-based approach. Learning programming is considered as a difficult course for the beginners, specifically for the first-year student. The opinions of 175 first-year undergraduate students at School of Computing, Universiti Teknologi Malaysia 2018/2019 session regarding their experience in the assessment of skill-based test 1 and test 2 were collected via an online survey. The result of classifying students’ opinions using the NB algorithm had a negative prediction accuracy of 92% and a positive prediction accuracy of 75%. NB had a prediction accuracy of 85% which outperformed both the SVM with 70% and lexicon-based approach with 60% accuracy. The result shows that NB works better than SVM and Lexicon-based approach on small dataset. The findings from the analysis of the survey show that the student’s sentiment is classified as negative, which implies that the skill-based test is difficult and gives scary emotions to the students which may further affect students interest in programming assessment. The key finding of this study discovers that the policy of awarding zero scores to students’ whose program did not compile successfully, hinders the programming assessment of first-year undergraduate students in the School of Computing, Universiti Teknologi Malaysia. |
|---|