论文部分内容阅读
针对葡萄酒物理和化学数据成分冗余,提出了两种葡萄酒分类的算法,分别是主成分分析K均值和主成分分析自组织神经网络算法.这两种算法对葡萄酒的物理化学成分进行了主成分分析,提取了主要的影响因素,将输入维数降低,再利用K均值和自组织神经网络算法分别对葡萄酒进行分类和比较.实验结果表明,PCA-K-means和PCA-SOM都具有较高的准确率,都有一定的使用价值和可操作性,并且PCA-K-means算法优于其它的算法.
According to the redundancy of wine physical and chemical data, two wine classification algorithms are proposed, which are the K-means principal component analysis and the principal component analysis self-organizing neural network algorithm, respectively.The two algorithms have the main components of the wine’s physicochemical composition The main influencing factors were extracted and analyzed, the input dimension was reduced, and the K-means and self-organizing neural network algorithm were used to classify and compare wines separately.The experimental results show that both PCA-K-means and PCA-SOM have higher Accuracy, have a certain value and operability, and PCA-K-means algorithm is superior to other algorithms.