ANALISIS SENTIMEN TERHADAP MOBIL LISTRIK MENGGUNAKAN METODE BERT DAN NER

Air pollution from fossil fuel-powered vehicles poses serious health risks. To reduce greenhouse gas emissions, electric vehicles (EVs) have emerged as a greener alternative. However, EV adoption in Indonesia still struggles, mainly due to low public acceptance. This study analyzes Indonesian public sentiment toward EVs and the key factors influencing it, using Natural Language Processing (NLP) with BERT for sentiment classification and Named Entity Recognition (NER) for identifying important entities. The BERT model performed well, with 71.71% accuracy, 83.56% precision, 71.71% recall, 75.59% F1-score, and a misclassification error of 28.29%, outperforming Naïve Bayes and LSTM. Sentiment analysis found that 48.45% of the public expressed negative sentiment, 30.60% neutral, and only 20.95% positive. NER identified influential factors including public events, opinions, company reputation, product quality, pricing, and location.These findings offer important insights for policymakers and industry players in designing strategies to boost EV adoption in Indonesia.