A horseshoe based prior for shrinkage towards a predefined parametric subspace.

• Paul Wiemann ^[1] ; Thomas Kneib
  1. [1] Göttingen University,
• Localización: Proceedings of the 35th International Workshop on Statistical Modelling : July 20-24, 2020 Bilbao, Basque Country, Spain / Itziar Irigoyen Garbizu (ed. lit.), Dae-Jin Lee (ed. lit.), Joaquín Martínez Minaya (ed. lit.), María Xosé Rodríguez Álvarez (ed. lit.), 2020, ISBN 978-84-1319-267-3, págs. 259-264
• Idioma: inglés
• Enlaces
  • Texto Completo Libro
• Resumen

  • We introduce a new prior hierarchy that allows shrinking of smooth spline-based functional e ects towards a prede ned vector space of parametric functions. Instead of shrinking each spline coecient towards zero, we adapt the horseshoe prior to control the deviation from the prede ned vector space.

    Furthermore, the prior presented regularizes the wiggliness of the estimated e ect.

    In this paper, we start with an application to energy consumption in Germany, then introduce the technical details and describe the prior's desirable shrinkage properties. We conclude with a simulation study to assess the validity of our approach.