Deducing neighborhoods of classes from a fitted classification model

• Alexander Gerharz ^[2] ; Andreas Groll ^[2] ; Gunther Schauberger ^[1]
  1. [1] Technical University Munich

     Technical University Munich

     Kreisfreie Stadt München, Alemania

     
  2. [2] TU Dortmund 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. 96-101
• Idioma: inglés
• Enlaces
  • Texto Completo Libro
• Resumen

  • For complex statistical or machine learning models, interpretable machine learning methods can be used to make up for the lack of interpretability.

    The method proposed here helps to understand the partitioning of the feature space into predicted classes in a classification model. Basically, it observes the changes of the predictions after slight manipulations of specific metric features.

    The observed changes can then be interpreted as neighboring classes in the feature space. An example is shown with the iris classification task.