Abstract:
Distance learning has become a necessary option to the education system for almost every nation because of the COVID-19 outbreak. Due to the closure of educational instit...Show MoreMetadata
Abstract:
Distance learning has become a necessary option to the education system for almost every nation because of the COVID-19 outbreak. Due to the closure of educational institutions, which creates obstacles to students' learning in the current COVID 19 pandemic environment, the role of information technology has gained momentum. With the current technology advancements, online learning is made possible for everyone. Although this method appears to be helpful to students and instructors, the effectiveness depended on several factors. For instance, the availability of internet services and the economic capacity of the users may affect the user experience of online learning. The student's reaction is desperately needed to enhance the university's effectiveness and acquire insight into student wants. However, it is difficult to collect and evaluate all the text data on social media, particularly Twitter. In this research, the analysis regarding perception students towards online learning is done through the Natural Language Processing technique which is sentiment analysis. The aim for this study is to determine the terms or keywords used for online learning and identify which machine learning classifiers work best with a large dataset. This research used a Twitter dataset that consisted of 38,602 tweets posted between 23rd July and 14th August 2020. Firstly, pre-processed the data to remove irrelevant tweets. Second, classify the tweet into three classes namely positive, negative, and neutral using rapid miner. Consequently, several machine learning classifiers are trained using different techniques which are Support Vector Machine (SVM), Naïve Bayes (NB), Decision Tree (DCT), and Random Forest (RF). Support Vector Machine classifier with a percentage split 80:20 using VADER Lexicon managed to get the highest accuracy which is approximately 90.41%. Finally, the result of the sentiment analysis has been shown using Power BI data visualization for better understanding. For fut...
Published in: 2022 3rd International Conference on Artificial Intelligence and Data Sciences (AiDAS)
Date of Conference: 07-08 September 2022
Date Added to IEEE Xplore: 26 October 2022
ISBN Information: