Social Media Sentiment Analysis of Thermal Engineering Students for Continuous Quality Improvement in Engineering Education / Wandeep Kaur ...[et al.]

In an academic institution, deciphering the opinions of students is the key that ensures the institution continues to strive within the education industry. Extracting implicit information from student opinions are vital in ensuring the standard of education continuously improves, ultimately leading to student retention and increase number of student intake within the institution. Sentiment analysis is a field of study that is interested in extracting sentiments from opinions extracted from written text. These techniques determine if an opinion is penchant towards positivity or negativity. The main aim of this paper is to conduct a preliminary analysis on the opinions of students taking Thermal Engineering (MEC551) from Universiti Teknologi Mara (UiTM) with regard to course tools. Data collected from Facebook was subjected to cleaning and pre-processing. A supervised machine learning algorithm was employed for sentiment classification purpose which was implemented using Rapid Miner. Algorithms were compared and results indicate Support Vector Machine (93.6%) outperformed Naïve Bayes (90.1%) and K-Nearest Neighbour (90.2%) in terms of accuracy and was able to correctly classify the text accordingly. This in return indicates students were very much interested in being able to interact and discuss on questions and queries via Facebook as well as address some fears they had related to exams and assignments seamlessly with their classmates as well as lecturer.