Aspect-based sentiment analysis towards technical and vocational education and training in Malaysia
Initiatives to improve public opinion towards technical and vocational education and training (TVET) have been increased by the government of Malaysia. However, to observe these sentiments with more transparent, analysis on public opinion is necessary. This research aims to assess the public sentime...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2019
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/96389/1/NurulAshikinAbdSamadMCS2019.pdf.pdf http://eprints.utm.my/id/eprint/96389/ http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:143443 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Initiatives to improve public opinion towards technical and vocational education and training (TVET) have been increased by the government of Malaysia. However, to observe these sentiments with more transparent, analysis on public opinion is necessary. This research aims to assess the public sentiment regarding to TVET in Malaysia by performing aspect-based sentiment analysis. This study took advantage of the data availability from social media where public nowadays tend to express their feelings towards any products and services. Twitter appears as one of the most common social media platforms in which, countless of users can participate and interact at any time. The data from Twitter are unstructured by nature thus further mechanism are needed to provide more meaningful information for future uses. A series of text pre-processing strategies were implemented in this study to improve the process of aspect extraction and classification. Topic modelling technique, Latent Dirichlet Allocation (LDA) was used to extract aspect category during aspect extraction process. The lexicon-based classifiers; SentiWordNet (SWN) and Valence Aware Dictionary and Sentiment Reasoner (VADER) and machine learning classifiers; Naïve Bayes (NB) and Support Vector Machine (SVM) were used to classify the tweets sentiments. The performance of the classifiers was observed based on the results of precision, recall, f-measure, and accuracy. The finding revealed that the public sentiment for five (5) identified aspects for TVET in Malaysia; Student, Course, Employability, Skill and Accreditation inclined towards positive sentiments. SVM shows the highest accuracy among other classifiers with an acceptable accuracy of 72%. The results from this study were expected to give beneficial insight for TVET stakeholders specially the governing bodies and TVET providers to plan for improvisation strategies. |
|---|