葡萄酒评价模型的设计与求?/p>
黄亚?/p>
,
韩磊,王梦瑶
(安徽师范大学数学计算机科学学院,安徽,芜湖?/p>
241000
?/p>
关键词:
T
检验;改进
K
均匀聚类分析;模拟退火;广义回归神经网络
?/p>
要:
针对葡萄酒质量评价的多样性和复杂性等问题,围绕评酒员评价的差异性、酿酒葡萄的分级、理化指标对?/p>
萄酒质量的影响分别建立模型,并对结果进行了详细的分析。首先采用置信区间法降低同一酒样的变异系数,按照?/p>
方差贡献率进行综合评分,在此基础上,通过组内评价指标变异度的检验,进行二次方差检验两组评分结果的可信度;
对于葡萄酒的分类,利用数据挖掘提取方法,得出主成分,为了改进
K
均值聚类算法的局限性,提高聚类的有效性;
最后考虑多维变量之间的关系,提出了基于广义神经网络模型,研究酿酒葡萄和葡萄酒的理化指标对葡萄酒质量的?/p>
响程度。并通过实际数据进行仿真,结果显示了提出的模型具有一定的合理性和有效性?/p>
中图分类号:
O235
文献标识码:
A
文章编号?/p>
(2013) 04
The design and solution of wine evaluation model
Huang Yakun
,
Han Lei,
Wang Yang.
(School ofMathematics& Computer Science, Anhui Normal University, Wuhu241000,China )
Key words
?/p>
T inspect; Improved K uniform clustering analysis; Simulated annealing; Generalized regression neural network
Abstract
?/p>
In this paper , we aim at the issue of quality assessment , and center on the Significant difference of the
tasting members
?/p>
evaluation results and reliability
?/p>
classification of wine grape
?/p>
the contact between the physical
and
chemical
indicators
of
the
wine
grape
and
wine
?/p>
the
effect
and
evaluation
of
the
physical
and
chemical
indicators of the wine grape and wine to the quality of the port wine , build models respectively and do a detailed
analysis of the result. Firstly we using the confidence interval method to reduce the coefficient of variation of the
same wine sample, in accordance with its variance contribution rate of the composite score, on this basis, we do
the
second
variance
test
two
sets
of
ratings
results
credible
by
group
evaluation
variability
inspection;
To
the
classification of wine, our data mining extraction method obtained the principal components, in order to improve
the
limitations
of
the
K-means
clustering
algorithm,
to
improve
the
effectiveness
of
the
clustering
simulated
annealing;
Finally,
we
consider
the
relationship
between
the
multi-dimensional
variables,
we
propose
training
relevant sample data based on generalized neural network model to study the impact of the physical and chemical
indicators of the quality of the wine in the wine grape and wine; simulation and actual data, the results show that
the proposed model has a rationality and effectiveness.
1
问题背景
目前,葡萄酒质量的鉴别主要靠感官分析和理
化指标分析的方法进行评价
[1]
,如确定葡萄酒质?/p>
时一般是通过聘请一批有资质的评酒员进行品评?/p>
每个评酒员在对葡萄酒进行品尝后对其分类指标打
分,
然后求和得到其总分?/p>
从而确定葡萄酒的质量?/p>
酿酒葡萄的好坏与所酿葡萄酒的质量有直接的关
系,葡萄酒和酿酒葡萄检测的理化指标会在一定程
度上反映葡萄酒和葡萄的质量?/p>
在此基础上,
本文
针对网上搜索的相关葡萄酒和酿酒葡萄的成分?/p>
据。从数学建模角度,讨论以下问题:一、分析两
组不同评酒员的评价结果有无显著性差异;二、根
据酿酒葡萄的理化指标和葡萄酒的质量对酿酒?
萄进行分级;三、研究酿酒葡萄和葡萄酒的理化?
标对葡萄酒质量的影响?/p>
2
模型假设与符号说?/p>
针对特定的背景,为了更好的进行问题说明,
给出以下假设和相关符号说?/p>
?/p>
1
)两组评酒员在对酒样进行评价的过程中不存
在明显偏好,评价总体较客观;
?/p>
2
?/p>
葡萄酒的质量客观上与酿酒葡萄的好?/p>
直接
有直接关系,主观上与评酒员的评分有直接关系;