Skip to main content

Goodness of fit and predictive ability in survival analysis

Objective



Research objectives and content
Survival analysis is a specific branch of statistics whose aim is modelling the time to an event. Examples of 'failure' time include the lifetime of machine components in industrial reliability, the duration of strikes or period of unemployment in economics, the survival time of patients in clinical trials. In any of the mentioned situations model checking is a crucial step of the analysis. The first objective of this project is developing suitable procedures for identification of anomalous observations (individuals who are poorly predicted by the final model). Further investigation of these subjects may be of great impact in medical research. As second goal we wish to develop measures of model predictive capability. though prediction is often the final task both in medicine and economics, this topic has received relative little attention in survival analysis. Thus a contribution in this direction can stimulate research in an open field.
Training content (objectives, benefit and expected impact)
A deeper knowledge in survival analysis can be acquired by collaborating with the research team of the host institution.
Links with industry / industrial relevance (22)

Coordinator

Universität Wien
Address
23,Spital Strasse 23
1090 Wien
Austria