CMStatistics 2018: Start Registration
View Submission - CMStatistics
B1085
Title: Clustering of spatially dependent functional data Authors:  Vincent Vandewalle - Inria (France) [presenting]
Cristian Preda - University of Lille (France)
Sophie Dabo - University of Lille (France)
Abstract: Two approaches for clustering spatial functional data are presented. The first one is the model-based clustering that uses the concept of density for functional random variables and logistic weights on the prior cluster probabilities depending on spatial coordinates. The second one is the hierarchical clustering based on univariate statistics for functional data such as the functional mode or the functional mean, and includes spatial weights in the distances computation. These two approaches take into account the spatial features of the functional data: two observations that are spatially close share a common distribution of the associated random variables. The two methodologies are illustrated by an application to air quality data.