Paper accepted at the European Conference on Information Systems (ECIS)
Our paper “GAM (e) changer or not? An evaluation of interpretable machine learning models based on additive model constraints” authored by Patrick Zschech, Sven Weinzierl, Nico Hambauer, Sandra Zilker, and Mathias Kraus has been accepted at the 30th European Conference on Information Systems in Timisoara, Romania.
In this work, we analyse, how a special family of interpretable machine learning models — named generalized additive models — performs across a variety of tasks. We compare these GAM models against state-of-the-art black-box models and find that GAM models can perform similarly well on multiple tasks.
Read the full article here.