Edgewise outliers of network indexed signals
This is an incremental improvement for network data analysis, addressing outliers in dependent graph-structured variables.
The paper tackles outlier detection in network-indexed multivariate data by introducing edgewise outliers and proposing a robust edgewise MCD algorithm, showing on simulated and real data that accounting for dependence structure improves detection.
We consider models for network indexed multivariate data involving a dependence between variables as well as across graph nodes. In the framework of these models, we focus on outliers detection and introduce the concept of edgewise outliers. For this purpose, we first derive the distribution of some sums of squares, in particular squared Mahalanobis distances that can be used to fix detection rules and thresholds for outlier detection. We then propose a robust version of the deterministic MCD algorithm that we call edgewise MCD. An application on simulated data shows the interest of taking the dependence structure into account. We also illustrate the utility of the proposed method with a real data set.