【摘要】 PVC(聚氯乙烯 )异型材的生产加工过程十分复杂 ,因此影响PVC异型材产品力学性能的因素众多。主成份分析法实质上是一种空间变换 ,它首先计算出原始样本集的协方差矩阵 ,再根据协方差矩阵的特征值对系统输出因子的贡献率进行取舍 ,保留下来的特征向量即为主成份向量 ,由这些主成份向量可构成新的样本集。此法有效消除了系统输入因子的相关性 ,使其数量减少 ,从而改善化工过程模型的结构更多还原

【Abstract】 The manufacturing process of PVC profiled materials is so complicated that there are many factors influencing the mechanic properties of PVC products. In fact, principal component analysis is a kind of spatial transformation. After obtaining the covariance matrix of the original sample set, we can accept or reject eigenvalues of the covariance matrix by its contribution rote to the output of the system. The accepted eigenvectors are called principal component vectors. The product of the original sample set and principal component vectors results in a new sample set. The method can eliminate the relativity and reduce the sum of input parameters and effectively improve the structure of chemical industrial process model.更多还原