红酒品种聚类分析

Cluster Analysis of Red Wine Varieties

摘要: 本文根据UCI红酒化学成分数据，指标为：红酒品牌、酒精浓度、苹果酸含量、灰度、灰的碱度、镁含量、总酚类化合物量、类黄酮量、原花青素年、颜色强度、色调、稀释酒、脯氨酸，进行红酒品种的聚类分析。本文基于K-means聚类法、层次聚类法及Dbscan聚类法，对红酒化学成分数据进行聚类分析。首先对数据进行筛选，并选择适合的、有意义的，且适合聚类的指标，然后进行数据预处理，最后通过R语言实现红酒品种的聚类。通过聚类结果进行解释，给出红酒的品种类别与质量的好坏。

Abstract: In this paper, according to the chemical composition data of UCI red wine, the indicators are: red wine brand, alcohol concentration, malic acid content, gray scale, alkalinity of gray, magnesium content, total phenolic compounds amount, flavonoid amount, procyanidin year, color intensity, hue, diluted wine and proline, to conduct cluster analysis of red wine varieties. Based on K-means clustering method, hierarchical clustering method and Dbscan clustering method, this paper conducts clustering analysis on the chemical composition data of red wine. First, the data were screened and appropriate, meaningful and clustering indexes were selected. Then, data pretreatment was carried out. Finally, the clustering of red wine varieties was realized through R language. The classification and quality of red wine were explained by clustering results.