SVM神经网络在葡萄质量评价中的应用

Grape Quality Appraisal by SVM Neural Networks

随着葡萄酒需求的大量增加,葡萄酒的质量受到了越来越多的关注。一般地,葡萄酒质量采用感官品尝的结果来评定,但是它经常受到多种因素的影响,同时葡萄酒的质量又没有统一的标准,因此葡萄酒的质量评价体系的建立亟待解决,而酿酒葡萄的质量直接决定了葡萄酒的质量。就葡萄的理化指标与葡萄质量之间的关系展开了分析,并建立了以酿酒葡萄理化指标的主成分为输入,酿酒葡萄的质量等级为输出的多输入单输出的SVM神经网络专家模型,实现了酿酒葡萄质量的智能分类。结合专家打分的结果可以看出,该模型的分类效果比较理想。

With the great increase of demand for wine,the quality of wine has been paid more and more attention.In general,it is often using sensory taste to class the quality of wine,but it is influenced by some factors.At the same time,the quality of wine has no uniform standard,so the wine quality classification has become an urgent work.The quality of grape directly determines the quality of wine. In this paper,it discuss the relationship between the physical and chemical indicators and grape quality,model a SVM neural network of physical and chemical indicators of grapes for input,the quality level of grapes for output,and achieve intelligent classification of grape quality.With the result of scores of experts,it can be seen that the classification of the model is very satisfactory.