Scalability properties of a new forward search algorithm

Abstract

We propose some computational improvements for a robust method for the identification of atypical observations known as forward search, which is based on the idea of monitoring quantities of interest, such as parameter estimates and test statistics, as the model is fitted to data subsets of increasing size. We provide a recursive implementation of the procedure which exploits the information of the previous step and we demonstrate the computational advantages of the new approach using both synthetic and real data.