KOMBINASI LEXICON-BASED DAN MULTINOMIAL NAЇVE BAYES CLASSIFIER DALAM ANALISIS SENTIMEN ARTIS SONG JOONG KI SEBAGAI BRAND AMBASSADOR SCARLETT WHITENING

Suci Kurniawati (Departemen Statistika, Fakultas Sains dan Matematika, Universitas Diponegoro)
Suparti Suparti (Departemen Statistika, Fakultas Sains dan Matematika, Universitas Diponegoro)
Sugito Sugito (Departemen Statistika, Fakultas Sains dan Matematika, Universitas Diponegoro)
Hasbi Yasin (Departemen Statistika, Fakultas Sains dan Matematika, Universitas Diponegoro)

Jurnal Gaussian
Vol 15, No 1 (2026): Jurnal Gaussian

Twitter as a social media platform can be utilized as a means of exchanging information about current events. The topic of K-Drama is often discussed by the public on Twitter and reaches trending topics, especially during the pandemic in Indonesia. The popularity of K-Drama in Indonesia has led to marketing strategies where actors are chosen as brand ambassadors. Song Joong Ki is one of the actors who has been chosen by Scarlett Whitening products to become their brand ambassador. The public expresses their responses on Twitter, and sentiment analysis is necessary to classify these responses as positive, neutral, or negative. The sentiment analysis combines the Lexicon-Based method and Multinomial Naїve Bayes Classifier. SentiWordNet is used in the Lexicon-Based classification method. The data preprocessing stage of this research includes cleansing, case folding, word normalization, tokenizing, filtering, and stemming. The combination of the Lexicon-Based method and Multinomial Naїve Bayes Classifier yielded an accuracy score of 81.50%. The words “jadi”, “brand”, and “ambassador” dominate the word cloud, indicating that the public extensively discusses the appointment of Song Joong Ki as the brand ambassador for Scarlett Whitening.