Sentiment Analysis of Government Policy Management on the Handling of Covid-19 Using Naive Bayes with Feature Selection

The government’s policy in dealing with the Covid-19 pandemic that has entered Indonesia has generated a response from the public, including Twitter social media users. Responses or comments from the community are also very religious, ranging from positive responses in supporting government policies in dealing with this Covid-19 to negative reactions in the form of criticism of the government, which is considered to have underestimated this virus. We use sentiment analysis to determine whether an opinion or comment containing a statement is positive or negative. The comments are rated by Indonesian language experts. In this study, the sentiment analysis process uses the TF-IDF method for word weighting, the information gain method for feature selection, and the Naive Bayes method for classifying public opinion. The results of this study indicate that the Naive Bayes algorithm is quite good in the text data classification model, with the highest accuracy, precision, and recall levels of 87.0%, 89.0%, and 98.0%, respectively. The average accuracy, precision, and recall levels are 86.2%, 87.6%, and 97%, respectively. This research can provide accurate and valuable information for the community regarding the response of Twitter users to the policy of handling Covid-19 in Indonesia.