<![CDATA[Government policies related to quarantine have generated various responses from the community, some people feel that The quarantine must be done so that the spread of the COVID-19 disease can be suppressed, but others also feel that this is detrimental to the community because their activities are being limited, this response can be found in their Twitter post. By analyzing the sentiments on people's Twitter posts, we can conclude whether a policy tends to get more positive or negative responses to the affected community. To carry out this analysis, deep learning method is used, namely Long-Short Term Memoryf (LSTM) with the addition of Word Embedding to 1364 independently crawled Indonesian people's Twitter data. Performance using the LSTM method produces 81% accuracy, 80% precision, 80% recall, and 81% f-measure. This LSTM method produces better performance than the other 2 methods, namely Naive Bayes and Recurrent Neural Network (RNN) with a difference of + 8%, with details of 74% accuracy, 72% precision, 74% recall, and 69% f-measure for the Naive Bayes method and 71% accuracy, 71% precision, 72% recall, and 72% f-measure for the RNN method.]]>
Copyrights © 2021
