A0301
Title: Inference for semi-structured regression
Authors: David Ruegamer - LMU Munich (Germany) [presenting]
Abstract: In modern data analysis, neural networks offer significant flexibility, particularly for large and non-tabular data. Semi-structured models capitalize on this flexibility by defining a regression model with an additive predictor that combines a structured effect, such as a linear effect, with an unstructured effect represented by a neural network. In this talk, we discuss options to derive inference statements for the structured model component when used together with an approximate uncertainty measure for the unstructured part in a semi-structured model.
