摘要
考虑到工业生产数据具有按工作点聚类和迁移的特点,提出了一种基于自适应仿射传播聚类(adAP)的多最小二乘支持向量机(LSSVM)算法进行软测量建模。该方法用adAP算法对训练样本进行分类以找到最优的聚类结果,采用LSSVM算法对各类样本分别建立子模型,并根据当前工作点所属子类的模型进行预测输出。将该方法用于聚丙烯熔融指数的软测量建模,结果表明,与其他方法相比该方法具有更高的回归精度和良好的泛化能力。
Since the industrial production samples are clustered around different operat- ing points, a soft-sensing method with multiple models based,on Adaptive Affinity Propagation Clustering Algorithm （adAP） and Least Square Support Vector Machine （LSSVM） is proposed. Classify the training samples into several classes using the adAP clustering to find the best clustering result, and train the sub-models by LSSVM accord- ing to corresponding sub-class samples. The test samples are assigned to appropriate sub-class, then predicted outputs are estimated by corresponding sub-models. The sim- ulation results of Melt Index indicate that the proposed method has better prediction ac- curacy and generalization performance.
出处
《青岛科技大学学报（自然科学版）》
CAS
北大核心
2012年第5期515-519,共5页
Journal of Qingdao University of Science and Technology:Natural Science Edition
基金
国家自然科学基金项目(51104175)
山东省自然科学基金项目(ZR2011FM014)
关键词
软测量
多模型
自适应仿射传播聚类算法
最小二乘支持向量机
soft-sensing, multiple model, adaptive affinity propagation clustering algo-rithm, least square support vector machine
