Proposals for project/master-thesis topics (Høst 2025/Vår 2026)
I am interested in Bayesian inference with application to biostatistics, in particular statistical methods for epidemiological data,
computer-intensive methods, spatial and spatio-temporal methods, and measurement error/spatial misalginment analysis. I am using mostly INLA or Stan for full Bayesian inference
Potential topics for project/master theses are:
- Adjusting for the survey design in the application of age-period-cohort models to complex survey data: Analyses of data from complex surveys often ignore the survey design, that means the fact that data are not collected completely at random. This may lead to biases in parameter estimates and underestimation of uncertainty. In this project we would like to investigate how the survey design could be incorporated in Bayesian age-period-cohort (BAPC) models. We will consider data collected from the National Health Interview Surveys (NHIS) in the US. Here, survey weights and information about the survey design are provided. Using real data analysis and possibly simulation studies we will compare the performance of standard BAPC models and BAPC models adjusting for the survey design in terms of paramter estimates and prediction. Ideas from Mercer et al. (2015, Annals of Applied Statistics, ,see here, will be borrowed. Project/Master-theses will typically combine methodological work, their implementation and application to simulated and/or real world data. Recommended courses in this aspect are TMA4300 Computer intensive statistical methods and TMA4315 Generalised linear models. I hope you find the project ideas outlined above interesting. Please contact me if you are interested in any of those topics or have any questions. I am also open to the suggestion of alterantive topics related to the fields outlined above.