Maximizing risk prediction by synthesizing multiple data sets accommodating systematic missing data and heterogeneous scales
Year: 2022
Type: dissertation
Abstract: Statistical analyses highly depend on the data quality and sample size. For lots of hypotheses, it is not sufficient to use only one data source. But data from different sources are often on a different scale and miss-aligned with the main data. Methods for treating this defiance additionally to high missingness and rare events were investigated in this thesis and shown in three studies from different study areas: prostate cancer, forestry, and fatal mountain accidents.
Institution: Technical University of Munich
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7067474282
