Text Mining Analysis of Changes in Consumer Perceptions of Food Index

This study conducted a topic modeling analysis with the BERT model to identify the themes of the review texts on seafood written by consumers and compare them with the empirical analysis results from surveys. More specifically, this study analyzed the relationship between consumers’ perceptions of seafood and the food index level, a subdomain of the PKNU Maritime Index. In particular, the correlation was explored to verify the relationship between the Korean consumers’ evaluation of seafood concerning taste, nutrition, variety, price, and safety and the food index. The findings were as follows. First, consumers’ perceptions of seafood were found to have a significant positive correlation with the food index. Second, the multiple regression analysis showed that, apart from the variety, the food index had a significant positive effect in the order of safety, nutrition, taste, and price. Such research findings showed that it is necessary to understand consumer perceptions of seafood, and the food index centered on the Northeast Asian Sea could be used as a foundation to improve consumer experience and satisfaction. This study further expanded the scope of existing marine humanities research by conducting text-mining analysis, providing theoretical and practical implications.