摘要
为了比较基于灰色欧几里德关联度的聚类方法与基于邓氏灰色关联度的聚类方法在煤种聚类中的应用效果,选取5种不同煤种的25个煤样为研究对象,测量其发热量、灰分、挥发分和硫分,并以此为基本参数,按照两类聚类方法的基本原理实现了所选煤样的聚类。结果表明,基于灰色欧几里德关联度的聚类方法的准确度为68%,基于邓氏灰色关联度的聚类方法的准确度达96%。
In order to make a contrastive study on two gray clustering approaches, the clustering analysis approaches based on gray Euclid relative grade and Deng＇s gray relative grade, for clustering coal samples, twenty-five coal samples with 5 different kinds of coal are chosed, four parameters, the calorific value, ash, volatile and sulfur content of 25 coal samples are measured, the chosed coal samples are clustered by the clustering approaches based on gray Euclid relative grade and Deng＇s gray relative grade with the four parameters are basic parameters. The results indicate that there are 8 error samples among 25 coal samples for the clustering approach based on gray Euclid relative grade, the precision is 68%, and there is only one error sample among 25 samples. For the clustering approach based on Deng＇s gray relative grade, the precision is 96%.
出处
《材料导报》
EI
CAS
CSCD
北大核心
2010年第22期80-83,共4页
Materials Reports
基金
广东省工业攻关项目(2005B10301051)
关键词
煤
聚类方法
灰色欧几里德关联度
邓氏灰色关联度
coal, clustering analysis approach, gray Euclid relative grade, Deng＇s gray relative grade
