摘要
为充分挖掘降雨量信息,通过分类模型,可以数据库中的数据项映射到摸个给定的类别中,进行初步趋势预测。回归分析可根据研究对象和目的,通过建立回归模型确定变量之间的依存关系和控制自变量来进行估计和预测。以昆明市西山区28个监控点降雨量数据来应用回归分析和聚类分析,比较系统聚类方法——Ward法聚类和K-mean方法的不同,结果显示在处理该批数据的这两种聚类方法中,以Ward法更为理想。得到的聚类结果有明显的地理位置聚拢性,基本符合实际情况。
In order to fully excavate the rainfall information,the data items in the database can be mapped to a given category,and the preliminary trend prediction is carried out.Regression analysis can estimate and predict the dependent relationship and control independent variables of variables based on the study object and purpose.28 in the mountain west conference in kunming,rainfall data from point to point application of regression analysis and cluster analysis,the difference between the system clustering method—the Ward method and method of K-scheme is compared,the results show that in dealing with the data of the two kinds of clustering method,the Ward method is more ideal.The clustering results obtained have obvious geographical clustering,which is basically in line with the actual situation.
作者
昌霞
刘赛娥
CHANG Xia;LIU Saie(Yunnan Land and Resources Vocational College,Kunming 650000)
出处
《计算机与数字工程》
2019年第8期2002-2005,共4页
Computer & Digital Engineering
关键词
聚类
回归
Ward法
clustering
regression
Ward method
