文章基本信息

标题：Multi-scale Clustering of Functional Data with Application to Hydraulic Gradients in Wetlands
本地全文：下载
作者：Mark C. Greenwood ; Richard S. Sojda ; Julia L. Sharp 等
期刊名称：Journal of Data Science
印刷版ISSN：1680-743X
电子版ISSN：1683-8602
出版年度：2011
卷号：9
期号：3
页码：399-426
出版社：Tingmao Publish Company
摘要：A new set of methods are developed to perform cluster analy-sis of functions, motivated by a data set consisting of hydraulic gradientsat several locations distributed across a wetland complex. The methodsbuild on previous work on clustering of functions, such as Tarpey and Ki-nateder (2003) and Hitchcock et al. (2007), but explore functions generatedfrom an additive model decomposition (Wood, 2006) of the original time se-ries. Our decomposition targets two aspects of the series, using an adaptivesmoother for the trend and circular spline for the diurnal variation in theseries. Dierent measures for comparing locations are discussed, includinga method for eciently clustering time series that are of dierent lengthsusing a functional data approach. The complicated nature of these wetlandsare highlighted by the shifting group memberships depending on which scaleof variation and year of the study are considered.
关键词：Cluster analysis; functional data analysis; generalized additive;model; wetlands; hydrology; groundwater.