Decoding Customer Voices: Sentiment-Based Evaluation of Nasi Kuning Karim

Abstract

This study aims to analyze customer sentiment toward Nasi Kuning Karim, which is a small traditional food business in Pontianak. Customer reviews were collected from online food delivery platforms and analyzed using a Naïve Bayes Classifier, with N-gram feature extraction using unigram, bigram, and trigram models. Model performance was evaluated using accuracy, precision, recall, F1-score, and classification error. The results show that the unigram model achieved the highest accuracy of 78.38%, followed by the bigram model at 73.87%, and trigram model at 64.86%. The sentiment analysis indicates that customers often praise the taste and portion size as the business’s core strengths, while negative reviews highlight risks to customer trust. These findings suggest strategies in customer relationship management, quality control, and digital marketing to improve satisfaction and long-term sustainability.