Rough clustering of Korean foods based on adjectives for taste evaluation

摘要

There are many adjectives to express the taste of Korean foods. In the paper, those adjectives are used to categorize the foods based on rough tolerant relation as the same way as document clustering. Before clustering, we selected 87 adjectives that are frequently used for taste evaluation and obtained the sets of adjectives that can be used for the taste of 51 kinds of Korean cuisine and refreshments. For clustering non-hierarchical algorithm was used. The foods with similar ingredients, state of cuisine for serving, cooking methods and taking methods share the same cluster as we expected. The set of adjectives corresponding cluster representative can be used as linguistic scales for evaluating taste of foods in the category.