Sentiment Analysis of Customer Review Behavior in the Online Market toward Health Supplement Products

This research analyzed customer sentiment through review behaviors towards health supplement products in the online market by utilizing natural language processing techniques. Review data were collected from 10 brands of health supplements, comprising 6,284 reviews in Thai, on the Shopee platform from July 2023 to March 2024. The analysis results revealed that positive sentiment scores higher than 0.6 clearly reflect customer satisfaction. K-Means and Fuzzy C-Means were employed for clustering models to enhance accuracy. To determine the optimal number of clusters, Elbow Method and Silhouette Score were utilized. Ultimately, it was found that customers were divided into two clusters (K=2), with key focus on words such as "efficacy," "product," "good," "taste," "consume," and "quality," which are critical factors influencing purchasing decisions. Furthermore, an evaluation of clustering quality using the K-Means and FCM models showed no significant differences. As indicated by the Silhouette Score (0.5319) and Davies-Bouldin Index (0.6991), both models held comparable clustering performance. This study offered important business implications, including customer segmentation, behavior analysis, marketing strategy development, and product/service improvement to help strengthen customer relationships and enhance competitiveness in the rapidly changing online market. Consequently, the findings of this research served as a vital tool for forecasting trends, developing effective strategies, and meeting future customer needs.