SHSADReM

Structured additive distributional regression provides a unifying framework for regression models to overcome some of the limitations of common regression specifications, namely instead of focusing on the expectation of the response, they take on a broader view and enable modeling the complete conditional distribution of the response in terms of covariates. In addition, one considers structured additive predictors that additively combine a variety of regression effects including nonlinear effects of continuous covariates, varying coefficients, spatial effects of various forms, random intercepts and random slopes, as well as a number of additional effects. This project focuses on extending and developing statistical methodology for structured additive distributional regression and on applying the developed methodology in timely and challenging research areas from the life sciences and social sciences.