Sentiment Analysis of User Feedback in e-Learning Environment
Mohd Asri Omar, Mokhairi Makhtar, Mohd Fauzi Ibrahim, Azwa Abdul Aziz · International Journal of Engineering Trends and Technology · 2020
Sentiment analysis (SA) is prevalent now; because it can yield useful insight from high-volume subjects and unstructured data mainly from social media networking sites and micro-blog websites or known as user-generated contents.SA in user-generated contents is difficult due to the informal nature of the communication.The informal nature introduces additional variables and properties that have to be evaluated compared to formal texts, necessitating additional resources spent on annotating the data and training the classifiers.We explore two most common methods in classifying user-generated contents called lexicon-based approach by using VADER Sentiment Analyser and Machine Learning (ML) approaches by using Naïve Bayes and Decision Tree classifiers.Our primary objective is to study the accuracy of the solutions and then apply the best solution to 126 students' feedbacks 126 student feedbacks toward an e-learning environment.The purpose is to extract the sentiment against it to acquire the initial picture of student's perception on the implementation of e-learning; so the effectiveness of its implementation can be improved.The data pre-processed and then analysed using Python as programming tool.The results show VADER outperformed two selected ML classifiers, it can achieve approximately 90% in accuracy.From the results, we conclude that VADER sentiment analyser was doing well and better than ML in SA toward user-generated content.The results on the e-learning environment also suggest further analysis should be done towards this e-learning platform to complete this initial study.