STUDI EKSPERIMEN METODE NAIVE BAYES CLASSIFIER PADA ANALISIS SENTIMEN PENGGUNA TWITTER TERHADAP OPINI PERPINDAHAN IBU KOTA

The president's official decision regarding the relocation of the capital city, which previously was in Jakarta, will be moved to East Kalimantan, will be a big thing and cause a lot of debate from both parties who agree and disagree because this project will require a lot of funds and will have an impact on various sectors. So with this research, it can be used as an information on the new capital relocation program, with sentiment analysis, information on topics discussed by the community of several Twitter users can be obtained. This study uses 1290 tweet data consisting of 686 negative sentiments and 603 positive sentiments which are used as training data to create evaluation models using confusion matrix and split validation. In this study, using TF-IDF word-weighted feature extraction with the Naive Bayes method. The result from the experiments carried out the best accuracy in split data 90:10 with 76.74%. The model that has been made is implemented in 1115 test data resulting in 799 negative sentiments and 316 positive sentiments. From the test data, it produces several words that appear, including negative sentiment, namely negara, rakyat, proyek, covid, corona, and pandemi. From these words can be described that its better for the goverment to take care more during the rainy season like a flood and disasters that are hitting the whole world, namely the corona or covid 19 virus, while positive sentiment results in words that often appear including Kalimantan, Jakarta, bangun, mulai, and Jokowi. These words illustrate that from the people who provide positive sentiment support the process of moving the new nation's capital to Kalimantan by Jokowi. It can be concluded that the results tend to be more of a negative sentiment