COVID-19 Social Safety Nets Sentiment Analysis on Twitter Using Gated Recurrent Unit (GRU) Method

ABSTRACT

One of the Indonesian government's programs in dealing with Covid19 problems is the Social Safety Net program which is given to the community, especially Covid19 assistance which is given every month to the community. Based on the assistance provided by the government, many people expressed their opinions through Twitter social media. This study aims to analyze the sentiment on Twitter tweets regarding the Social Safety Net Program from March to December 2020. The data collected is 4061 tweets data. The data is classified into two classes, namely positive and negative. The classification algorithm used is Gated Recurrent Unit (GRU). Hyperparameter testing is carried out in order to produce an optimal model. In the optimal GRU hyperparameter, when there are 10 GRU units, the activation function is sigmoid, the optimizer used is Adam, the batch size is 128, with 10 epochs of iteration and 0.2 dropout size. The GRU model produces an f1score of 92.09%, a precision of 90.34%, and a recall of 93.90%. © 2021 IEEE.