A Study on Consumer Perception of Decaffeinated Coffee Using Text-Mining Analysis

This study aims to understand consumer perception and trends related to decaffeinated coffee through text mining text analytics), identifying key keywords connected to decaffeinated coffee. From March 14, 2019, to March 13, 2023, text documents on the network were collected, and the collected data underwent refining processes using Textom. A total of 14,204 pieces of data were collected, and after preprocessing to remove unnecessary data, a refined dataset of 7.7 MB was obtained. The collected and refined data were analyzed using natural language processing for frequency analysis, co-occurrence frequency analysis, revealing key keywords. The relationships between keywords were analyzed through connection structure and centrality analysis. Based on this, cluster analysis was conducted, identifying four core keyword clusters with structural similarity. Sentiment analysis on consumer emotions regarding decaffeinated coffee was conducted using the sentiment dictionary provided by Textom. The results of text mining analysis are as follows. Firstly, it was observed that 'franchises' are prominently discussed in decaffeinated-related documents. Secondly, centrality analysis for network connection structure revealed a close association between decaffeinated and coffee, followed by a strong connection to franchises. Centrality analysis was used to understand the network structure of keywords related to decaffeinated coffee, which were primarily associated with the attributes, taste, and purchase intentions of decaffeinated coffee. Thirdly, through CONCOR analysis, four key groups were identified, focusing on consumer experiences with decaffeinated coffee, health-related concerns, interest in new products, and personal transactions. Particularly, health-related keywords significantly influenced consumer concerns regarding caffeine