CMStatistics 2020: Start Registration
View Submission - CMStatistics
Title: Additive models: Smooth backfitting, high dimensions, random mixed forests Authors:  Enno Mammen - Heidelberg University (Germany) [presenting]
Abstract: The purpose is to report on recent developments in the study of nonparametric additive models where the expectations of response variables are modeled as the sum of nonparametric functions of covariables. We will discuss extensions and modifications of the model. We will consider the high-dimensional case where the number of additive components converges to infinity and sparsity assumptions are made. We will discuss random mixed forests, a modification of random forests designed for additive models, and nonparametric ANOVA type regression models.