Enhancing Restaurant Success Insight through NLP-Based Topic Modeling and Emotion Analysis of User Reviews - Dalam bentuk buku karya ilmiah

FADHILAH NURDZATI

Informasi Umum

Kode

25.05.918

Klasifikasi

000 - General Works

Jenis

Karya Ilmiah - Thesis (S2) - Reference

Subjek

Data Mining

Dilihat

98 kali

Informasi Lainnya

Abstraksi

<p>In the highly competitive culinary industry, especially in fast-growing markets like Bandung, understanding customer preferences and experiences is essential for business survival. This study examines the characteristics that contribute to the success of restaurants and cafes by analyzing user- generated content from TripAdvisor reviews. Using the PERVAL framework, we identify three dimensions of perceived value, such as emotional, economic, and functional, that influence customer loyalty. Topic modelling is conducted using BERTopic to extract key themes from reviews, while sentiment classification applies Plutchik’s Wheel of Emotions to uncover emotional tones in customer feedback. The findings reveal that emotional value, particularly related to ambiance, comfort, and overall dining experience, is most frequently mentioned and strongly associated with positive sentiments such as joy, love, and optimism. In contrast, economic and functional values are less emphasized in positive reviews. These insights highlight the importance of creating emotionally satisfying experiences to enhance customer satisfaction and loyalty in the culinary industry. The study contributes to the development of the PERVAL framework by integrating advanced natural language processing tools in analyzing online reviews. Strategic recommendations are presented for culinary business owners to improve service quality, curate memorable experiences, and strengthen customer loyalty in a competitive and digitally influenced marketplace. This paper extends the PERV AL framework through the integration of natural language processing and emotion classification, offering a novel approach to understanding customer value in the context of online reviews.</p>