Penerapan Algoritma Cerdas Bidirectional Encoder Refresentations From Transformers Dalam Menganalisis Opini Publik Terhadap Produk Yang Mengalami Boikot

Social media, particularly Instagram, has become a primary platform for expressing opinions and participating in public discussions on various social, economic, and political issues. One of the prominent issues on social media is product boycotting. Boycotting a product can significantly impact the brand's image and sales. Famous brands such as McDonald’s, KFC, Starbucks, Burger King, and Pizza Hut are the main targets in boycott actions. This study uses a dataset of 1,750 comments from Instagram accounts on related products. The data is divided into two labels, positive and negative, based on automatic labeling from transformers and manual labeling. Sentiment analysis results show that McDonald’s has 41.43% positive sentiment and 58.57% negative sentiment, KFC has 85.14% positive and 14.86% negative, Starbucks has 97.71% positive and 2.29% negative, Burger King has 50% positive and negative, and Pizza Hut has 80.57% positive and 19.43% negative. Modeling results using the pre-trained Bidirectional Encoder Representation From Transformers (BERT) from Bert-Base-Uncased show accuracy results for McDonald’s products at 84.14%, KFC products at 95%, Starbucks products at 94.16%, Burger King products at 91.42%, and Pizza Hut products at 93.80%.