Title

Riding Down the Bay: Space-Time Clustering of Ecological Trends

Document Type

Article

Publication Date

8-1-2018

Department

Political Science, International Development, and International Affairs

Abstract

The Chesapeake Bay Program, initiated in 1983, is a regional partnership between several state governments, federal agencies, and advisory groups that is involved in the cleanup and restoration of the Bay. To study the ecological trends in the area, we propose a new data‐driven procedure for optimal selection of tuning parameters in dynamic clustering algorithms, using the notion of a stability probe. We refer to the new procedure as Downhill Riding (DR) because of the dynamics of the clustering stability probe. We study the finite sample performance of DR when clustering benchmark Iris data and synthetic times series, and illustrate the methods using data on water quality in the Chesapeake Bay.

Publication Title

Environmetrics

Volume

29

Issue

5-6

Find in your library

Share

 
COinS