基于奇异值分解和K means算法的近海水质数据分类方法  

Clustering Method of Water Quality in Coastal Water Areas Based on Singular Value Decomposition and K Means Algorithm

作　　者：张丽梅[1] 宛立 

机构地区：[1]大连海洋大学理学院,辽宁大连116023 [2]海洋与土木工程学院,辽宁大连116023

出　　处：《数学的实践与认识》2014年第20期140-147,共8页Mathematics in Practice and Theory

基　　金：辽宁省科技计划项目(2012216012)

摘　　要：以辽东湾某生态监测区水质监测数据为例,以矩阵的奇异值分解和K means算法为分类工具,给出生态监测区水质监测数据的分类方法.方法具有以下特点:通过奇异值分解简化并加速了类比过程,通过动态设置类K避免了K means算法先设定类数的不足,还探讨了对少量新增监测数据的归类问题.方法对近海海水水质监测数据分类具有普适性.Taking an example of the monitored data of water quality in some ecological monitoring area in LiaoDong Bay.By means of the singular value decomposition and K means algorithm as the classification of tools,the clustering method of the monitored data of water quality in some ecological monitoring area is presented.This method has the following characteristics.The process of analogy are simplified and accelerated through the use of the singular value decomposition.The class number K is set up dynamically and this makes up for the shortage of fixed class number in advance.The classification problem for a small amount of new monitoring data is discussed in this paper.The method has universality for clustering the offshore sea water quality monitoring data.

关 键 词：海水水质 奇异值分解 K MEANS算法 正交 

分 类 号：X834[环境科学与工程—环境工程]