Sparse spatial pattern selection via component-wise gradient boosting

Tobias Hepp, Friedrich-Alexander-Universität Erlangen-Nürnberg

Co-authors: Anna von Plessen, Georg-August-Universität Göttingen; Nadia Müller-Voggel, University Hospital Erlangen; Elisabeth Bergherr, Georg-August-Universität Göttingen

Abstract: In this work, we present a new approach to sparse spatial pattern estimation based on regularization via gradient boosting.

A first version of the algorithm is evaluated in a scalar-on-function regression framework motivated by a neurophysiological research question on tinnitus.