Hierarchical Spatio-Temporal Change-Point Detection

Detecting change-points in multivariate settings is usually carried out by analyzing all marginals either independently, via univariate methods, or jointly, through approaches. The former discards any inherent dependencies between different and the latter may suffer from domination/masking among of distinct marginals. As a remedy, we propose an approach which groups with similar temporal behaviors, then performs group-wise change-point detection. Our based on hierarchical clustering using distances adjust for dependencies. Through simulation study show that our approach, preventing domination/masking, significantly enhances general performance employed detection method. Finally, apply to two datasets: (i) Land Surface Temperature Spain, during years 2000–2021, (ii) WikiLeaks Afghan War Diary data.