Using DAGs to identify the sufficient dimension reduction in the Principal Fitted Components model

Publication date: Available online 27 August 2018Source: Statistics & Probability LettersAuthor(s): María Eugenia Szretter NosteAbstractWe identify the sufficient reduction for the Principal Fitted Components model under mild conditions which generalize those considered in previous works. We give a short proof of the main result based on the d-separation for directed acyclic graphs (DAGs), linking two areas that, to our knowledge, have not been linked before in statistics.

Source: Statistics and Probability Letters - August 28, 2018 Category: Statistics Source Type: research