Reviewing the effectiveness of lexicon-based techniques for sentiment analysis in massive open online courses

Massive open online courses (MOOCs) platforms now play a big part in online learning. These platforms typically solicit course-related feedback from the community of registered students. The platforms and the students benefit from the analysis of the reviews. Any type of review may often be classified using rule-based, machine learning, lexicon-based, and hybrid approaches. The reviews of massive open online courses are categorized into good, negative, and neutral in this study using lexicon-based techniques such as Afinn, TextBlob, Vader, and Sentiwordnet. The reviews of the 234 courses available on the Coursera platform are used to conduct experiments. To compare the categorization results from lexicon-based techniques, the cosine similarity metric is utilized. Statistical measurements are also used to gauge how accurate the categorization is. The results section follows with a comparison analysis of the lexicon-based techniques. The results demonstrate that Afinn and TextBlob classifiers perform more effectively than Vader and Sentiwordnet classifiers.