【摘要】 鉴于主元分析法的降维特性和模糊C均值聚类算法良好的分类性能 ,本文在丙烯腈反应器操作参数的优化中 ,结合这两种方法 ,将主元分析处理后的数据作为新的样本输入 ,利用模糊C均值聚类算法进行优化操作。在保留原有信息的基础上 ,去除了冗余数据 ,加快了聚类速度。实验表明 ,混合算法的聚类结果比单纯的基于聚类优化的方法能较好地对操作参数的优化起指导作用。更多还原

【Abstract】 Because the principal component analysis (PCA) can drop dimensions fuzzy C-means cluster (FCM) offers good classifying capability, the author combines these two methods for the operation optimization of the acrylonitrile reactor. The data processed by PCA become the new inputs of FCM, then optimized by using FCM. On the basis of reserving the original data, the method not only removed the redundant data, but also made the speed of clustering faster. The result of this combined algorithm is better than that of pure cluster, and guiding optimization of operational parameters.更多还原