Abstract
Based on dietary records from 544 persons—female university students and their parents—we developed a database using 42,508 dishes consumed on 2,877 days. We then analyzed the database using the names of the dishes, the weights of foods used in them, and their nutrient values. Focusing on nutrient values and weights in selected food groups, we classified the dishes into 11 pattern groups by cluster analysis, in which protein, lipid, carbohydrate, weight of vegetables, fruit and milk as food groups were selected as variables in the model. There were three different type of large groups: “combined dish group”, “single dish group” and others. The “combined dish group” had three pattern groups: “compound staple food” (n=1,364; e.g. curry and rice), “compound main dish” (n=448; e.g. Sukiyaki), and “compound side dish (n=695; e.g. Oden). The “single dish group” had 6 pattern groups: “staple food” (n=5,916) containing 60 g of carbohydrate, “main dish” (n=1,789) containing 20.3 g of protein, “side dish” (n=4,226) with 70 g of vegetables, “milk and dairy products” (n=1,362) with 187 g of milk, and two separate “fruits” groups with 100 g (n=1,582) and 230 g of fruits (n=343) on average in each group. However, the majority of dishes were not classified into these 9 groups, but into “small main dish” (n=5,865) mostly served in lunch boxes or small pots, and “drink or small dish” (n=18,918). These 11 dish pattern groups have significant and important characteristics for dietary assessment and nutrition education in terms of nutrient content or food type.
