With the proliferation of digital platforms, online reviews have emerged as a representative form of digital content directly produced by consumers, as well as a core information resource that influences consumer decision-making and corporate competitiveness. In particular, the restaurant industry places growing importance on analyzing consumer experience through review content, owing to the intangibility and experiential nature of its services. However, most existing sentiment analysis studies have focused on English-language platforms such as Yelp and TripAdvisor, leaving Korean-language review content and Naver Place relatively underexplored. Accordingly, this study utilizes restaurant review content from Naver Place to compare the performance of various machine learning-based sentiment analysis techniques and proposes an analytical model that predicts customer satisfaction based on the linguistic characteristics of review content. Based on the analysis results, implications for restaurant content management and platform service improvement are also presented. This study extends the academic scope of Korean-language review content analysis while providing a foundation for data-driven decision-making and content strategy in the restaurant industry.
목차
Abstract I. Introduction II. Literature Review 1. Review Content as a Digital Content Asset 2. Electronic Word-of-Mouth (eWOM) Theory 3. Sentiment Analysis and Aspect-Based SentimentAnalysis 4. Machine Learning Techniques for Sentiment Classification 6. Characteristics of Korean Natural Language Processing III. Methodology 1. Data Collection 2. Data Preprocessing 3. Analytical Models IV. Results 1. Descriptive Statistics 2. Comparison of Sentiment Classification Model Performance 3. Customer Satisfaction Prediction Model Results V. Discussion and Conclusion Reference