Covid-19 is a disease that has been declared a global pandemic since March 2020. One of the challenges in dealing with the current Covid-19 pandemic is the widespread doubts about the use of vaccines, even though vaccination is one of the most successful ways to deal with infectious disease outbreaks. Vaccine hesitancy can be observed, among others, from public sentiment or perception on social media, one of them is Twitter. The existence of social media can affect the absorption of information received by a person, in this case social media is also a medium for anti-vaccine propaganda which can result in a decrease in public confidence in the Covid-19 vaccine. This study aims to develop a classification model using the Support Vector Machine (SVM) method for sentiment analysis of Tweet related to the Covid-19 vaccine. Several previous studies have conducted sentiment analysis related to Covid-19, but this research specifically conducts sentiment analysis on the topic of the Covid-19 vaccine so that data preparation and model configuration will be different. This study also uses the Design Science Research Methodology (DSRM) for research as a whole before focusing on the use of the SVM method. The results of the study consist of an algorithm for creating data sets and a classification model for sentiment analysis that can be used to determine public perceptions of the issue of Covid-19 vaccination. This study also compares the use of unigram and bigram tokenization. Based on the results obtained, the average value of each aspect of the evaluation measurement is higher when the bigram tokenization is used. Although higher, the value obtained has an insignificant difference in the range of 0.6% - 0.7%. In the evaluation results using unigram and bigram tokenization, the highest scores for all aspects of measurement, namely accuracy, recall, f-measure, and precision were 84%.
Copyrights © 2021