计算机工程与应用 ›› 2016, Vol. 52 ›› Issue (2): 175-179.

• 模式识别与人工智能 • 上一篇    下一篇

葡萄酒的模糊C均值聚类及其最优聚类选择

黄春娥,张  洪,张景胜,孙明星,叶志伟   

  1. 北京联合大学 生物化学工程学院,北京 100023
  • 出版日期:2016-01-15 发布日期:2016-01-28

Fuzzy c-means algorithm and optimal numbers of clusters for wine

HUANG Chun’e, ZHANG Hong, ZHANG Jingsheng, SUN Mingxing, YE Zhiwei   

  1. College of Biochemical Engineering, Beijing Union University, Beijing 100023, China
  • Online:2016-01-15 Published:2016-01-28

摘要: 现有以品酒员的感官指标来评定葡萄酒质量的方法受品酒员主观因素影响较大,导致葡萄酒质量评定结果存在较大的不确定性。针对葡萄酒质量的评定问题,根据影响葡萄酒外观、口感和香气的主要理化指标,结合模糊C均值聚类分析、F统计量及其变化的显著性水平评定聚类有效性的方法,给出了葡萄酒的客观的分类,并结合具体葡萄酒理化指标的原始数据,利用MATLAB编程实现,验证该方法的有效性。该方法克服了已有基于模糊聚类的葡萄酒分类方法中的不足,客观地评价葡萄酒的质量,为葡萄酒质量的评定提供可选择的方法。

关键词: 葡萄酒, 理化指标, 模糊C均值聚类, F统计量

Abstract: Popular method of wine classification is to invite wine tasters to taste the waited classification wine and give scores. This approach is influenced greatly by the subjective reasons of the wine tasters, and leads to uncertainty appear. Based on the evaluation of wine classification, an objective method is given. This approach depends on the mainly physicochemical indexes of wines which decide surface, mouth feel and fragrance, fuzzy c-means algorithm, F-statistical quantity and its significance level. Original data of wine is used to check the validity of the method with MATLAB program. This approach overcomes the deficiency of the existed methods, objectively evaluates wine quality, and proposes an alternative for wine classification.

Key words: wine, physicochemical index, fuzzy c-means algorithm, F-statistical quantity