Akademik Veri Yönetim Sistemi
Food Science and Nutrition, cilt.13, sa.7, 2025 (SCI-Expanded)
The study analyzed 2022 Eurostat data on the daily consumption of fruit and vegetables (FV) in 31 European countries, categorized by educational level (ISCED 2011: 0–2, 3–4, 5–8). Consumption was categorized into “0 portions,” “1–4 portions,” and “5+ portions” for adults aged 15+. K-means clustering (an unsupervised machine learning algorithm used to categorize the data into clusters), discriminant analysis (IBM SPSS Statistics 27, JASP 16.4), and one-way ANOVA were applied to classify four different country groups. The robustness of these clusters was assessed using Box's M test (a test for equality of covariance matrices), with results indicating no significant difference (p > 0.05). K-means clustering grouped the countries based on similar FV consumption patterns, while discriminant analysis and ANOVA (F-test) validated the clusters and found significant differences between them. The results showed that “0 portions” and “1–4 portions” were influential in determining the clusters across all education levels (p < 0.001), while “5+ portions” showed no significant relationship. Further analysis by education level revealed that “0 portions” had the greatest association with lower levels of education (level 0–2; F(3,27) = 13.168, p < 0.001), while “1–4 portions” was more influential with higher levels of education (level 5–8; F(3,27) = 19.380, p < 0.001). Clustering focused on total FV consumption, not on specific portions. Countries with similar socioeconomic and educational characteristics clustered together, indicating common dietary habits. The study shows the role of education in determining FV consumption and highlights the regional differences in diets across Europe.