Sentiment analysis on the issue of COVID-19 vaccination using LSTM and fasttext

The rapid spread of covid-19 worldwide made WHO decide covid-19 was a global pandemic. This spread impacts the Economy, Education, Social, and others. In response to this, the Indonesian government took action to procure covid-19 vaccination. This vaccination action is one way to reduce the spread of the covid-19 virus. The procurement of vaccinations in Indonesia has caused responses and opinions on various social media, one of which is Twitter. Many Indonesians give their opinion or opinion on Twitter about the procurement of vaccinations. Sentiment analysis is a way of analyzing how the views of the Indonesian people about vaccination. This study aims to create a classification model to determine the sentiment analysis of the Indonesian people regarding the procurement of vaccination using the Long Short-Term Memory (LSTM) method. Several studies are discussing sentiment analysis using the LSTM method, but in this research, fasttext embedding will be used, combined with selecting the proper optimization function and learning rate. The number of datasets resulting from crawling from Twitter is 1268. The best results from all LSTM model tests were obtained using fasttext embedding with SGD learning rate schedule, namely 85% accuracy and 85% f1-score. The addition of fasttext embedding can increase the model’s effectiveness by 21%.