SVMR在部分稳定氧化锆稳定率预测中的应用  

Partially Stabilized Zirconia Stability Prediction Based on SVMR

作　　者：刘志刚[1,2] 刘明清[3] 汪小志[1] 岳晓光[4] 武曲 

机构地区：[1]南昌大学环境与化学工程学院,南昌330031 [2]河南省矿山信息化重点学科开放实验室,焦作454000 [3]国家环境保护部华南环境科学研究所,广州510655 [4]武汉理工大学资源与环境工程学院,武汉430070 [5]太原科技大学计算机科学与技术学院,太原030024

出　　处：《中国陶瓷》2013年第12期62-65,共4页China Ceramics

基　　金：南昌大学博士科研启动经费项目(06301043);河南省矿山信息化开放实验室资助项目(KY2012-01)

摘　　要：为了预测部分稳定氧化锆稳定率,讨论了国内外数据处理的相关研究方法,采用回归型支持向量机(Support Vector Machine for Regression,SVMR)方法应用到预测中去。首先,通过工程实例建立预测模型;其次,使用Matlab仿真了SVMR模型和BP(Back Propagation)神经网络模型;最后,完成了两种模型下的部分稳定氧化锆稳定率预测。预测结果表明,SVMR方法的性能优于BP神经网络方法,回归型支持向量机是预测部分稳定氧化锆稳定率的有效方法。In order to predict the stability of partially stabilized zirconia, the related domestic and foreign research methods for data processing are discussed; partially stabilized zirconia stability prediction method based on support vector machine for regression has been put forward. First of all, relevant data for test are selected from the project example; secondly, support vector machine for regression model and BP neural network model are created by using Matlab; finally, experiment simulation has been finished. The predicted results show that, support vector machine for regression is an effective method for partially stabilized zirconia stability prediction, and the performance of support vector regression is better than BP neural network model.

关 键 词：部分稳定氧化锆 稳定率 回归型支持向量机 BP神经网络 MATLAB 

分 类 号：TP391.97[自动化与计算机技术—计算机应用技术]