Classification of Public Opinion on Vaccine Administration Using Convolutional Neural Network

Vaccine is one of the interesting discussions that are often found in various timelines, both from social media and local scope. The covid vaccine is a topic that is often discussed because the virus is still not over until now. Many researches are discussing how to stop this virus and produce various types of vaccines that are claimed to be able to stop cases of Corona Virus infection. Astrazeneca and Sinovac are two of several types of vaccines circulating in several countries and have a good reputation even though it differ from one another. Indonesia as one of the countries with the highest death rate due to Covid-19 with a total daily death of ±2000 people make Indonesia vulnerable to the level of transmission. To overcome this problem, Indonesia uses Astrazeneca and Sinovac as an effort to prevent deaths from Covid-19. This study focuses on comparing public opinion responses to both types of Astrazeneca and Sinovac vaccines using data collected via Twitter with the text mining technique. The data obtained then is preprocessed so that the data can be used properly for further processing so as to produce the appropriate model. This study used the Convolutional Neural Network (CNN) method in the training process which has a good reputation in making deep learning models. From the process that has been passed with 2000 tweets from Indonesian users, 1347 tweets were ready to be used in the training process and resulted in 59.5% positive responses and 40.5% negative responses for the Astrazeneca vaccine. In addition, for the Sinovac vaccine, 53.3% of positive responses and 46.7% of negative responses were obtained.