BOBY ANDIKA PUTRA, - (2024) SENTIMENT ANALYSIS OF PRESIDENTIAL CANDIDATES OF THE REPUBLIC OF INDONESIA USING NA¨IVE BAYES CLASSIFIER AND SUPPORT VECTOR MACHINE. SENTIMENT ANALYSIS OF PRESIDENTIAL CANDIDATES OF THE REPUBLIC OF INDONESIA USING NA¨IVE BAYES CLASSIFIER AND SUPPORT VECTOR MACHINE, - (-). ---. ISSN - (Submitted)

Preview

Text BOBY ANDIKA PUTRA_TUGAS AKHIR.pdf Download (8MB) | Preview

Abstract

Indonesia, as a democratic country, will hold presidential elections in 2024. The issue of presidential candidates has been discussed since two years before the end of the term of office. Three presidential candidates are emerging, namely Anies Baswedan, Ganjar Pranowo, and Prabowo Subianto. This has resulted in various opinions from the public, including support, neutrality, or putting down certain parties. This study aims to analyze public sentiment towards Indonesian presidential candidates in 2024. Data was collected using a scrapping technique using Python within the last 5 months, starting from May 1 to September 21, 2023. The data was automatically labeled using the Indonesia Sentiment Lexicon (InSet). The labeling results showed that Anies Baswedan received 492 positive sentiments, 2343 negative sentiments, and 121 neutral sentiments. Furthermore, Ganjar Pranowo had 199 positive sentiments, 2608 negative sentiments, and 114 neutral sentiments. Meanwhile, Prabowo Subianto has 900 positive sentiments, 1746 negative sentiments, and 199 neutral sentiments. The data was then processed with Naïve Bayes and Support Vector Machine algorithms to perform sentiment classification. The results show that SVM has higher accuracy than NBC, with an average SVM accuracy of 89.24%, while NBC is only 83.79%. Despite the high accuracy, modeling with the SVM algorithm requires more training time than NBC on all three datasets.

Actions (login required)

View Item