摘要
针对反浮选过程的被控对象复杂、数学模型不确定以及控制要求高等特点,提出一种基于主元分析和模糊聚类的数据预处理算法。采用模糊C均值聚类算法得到聚类中心,进行线形回归从而对过程变量数据进行了预处理。主元分析法则用来进行辅助变量的选取和输入高维向量的降维简化,针对主元变量采用径向基函数网络建立了系统经济技术指标的预测模型。根据工业实际生产数据进行的模型校验和误差分析表明,能够满足浮选过程控制的精度要求。
Aimed at the characteristics such as the complication of controled object, uncertainty of mathematical model and high requirement of control at the process of anti-flotation, a data pretreatment algorithm based on principal component analysis and fuzzy C-means clustering for flotation process was proposed. Linear regression of clustering centers obtained by fuzzy c-means clustering algorithm was introduced to carry on data pretreatment. Principal component analysis was adopted to select the auxiliary variables and reduce and simplify dimensions of input vectors. Aimed at principal component variables, radial basis function network was adopted to set up the prediction model of eanomy and technology index in flotation process. Model verification presemed by using real operating data from industrial production indicates that the model' s precision is good enough to satisfy the request of floatation process control.
出处
《鞍山科技大学学报》
2005年第6期427-431,共5页
Journal of Anshan University of Science and Technology
基金
国家自然科学基金资助项目(60474058)
辽宁省教育厅高等学校科学研究项目(202193396)
关键词
数据预处理
模糊C均值聚类
主元分析
浮选过程
径向基函数网络
data pretreatment
fuzzy C-means clustering (FCM)
principal component analysis (PCA)
flotation process
radial basis function (RBV) network