Prosjekt- og masteroppgaver tilbudt av Sara Martino
My field of research is Computational statistics, and I have long experience in developing and working with the INLA software, see the R-INLA page.
The projects I offer are linked or motivated by an application. The goal can either be answer a concrete question from a dataset using some statistical model, or develop a new statistical method with an application as motivation. In this last case one could use simulated data instead of observed ones.
Below is a list of suggestions. Some are just ideas from collaborators so not yet very concrete. Get in touch if you find something interesting or if you have you own idea, so we can discuss possible projects together.
Health equity in Europe
In this project we analyze data from the European Social Service (ESS), a bi-annual repeated cross-sectional, pan-European, general social survey that collects data on health, social determinants, attitudes, beliefs, and behavior patterns of the European populations through eleven rounds of data collection since 2002.
Several research opportunities are open using this data source to answer questions around health equity in Europe with focus on temporal or spatio-temporal analysis. It will be important to adjust for survey design to avoid biased estimates. We will use general mixed effects models in a Bayesian framework.
This project is also linked to the CHAIN (https://www.ntnu.edu/chain) project. It opens for two students to collaborate. For more information contact Sara Martino or Andrea Riebler.
Modelling patient’s transition times during cardiac arrest
When a patient experiences cardiac arrest, numerous medical interventions are employed in an effort to restore life. The effectiveness of these interventions depends on factors such as the patient’s age, existing comorbidities, and the underlying cause of the arrest. These treatments often induce several changes in cardiac rhythm, ideally culminating in the return of spontaneous circulation (ROSC). During resuscitation, the patient may transition through four potential clinical states: pulseless electrical activity (PEA), asystole, ventricular fibrillation/ventricular tachycardia (VF/VT), or ROSC.
In this project we will analyse data from patient experiences cardiac arrest at St. Olav hospital in Trodheim with the aim to understand how to model transitions times when one wants to account for individual variability.
The project opens for the possibility for two students to collaborate. For more info you can contact Sara Martino or Thea Bjørnland.
Analysing bycicle accidents on road networks: ReCyCLIST
In the research project ReCyCLIST (https://www.toi.no/recyclist/), the Transport Economics Institute and Sørlandet Hospital collaborate to collect data on accidents on bicycles, electric scooters and other forms of micromobility. People who come into contact with the healthcare system after injuring themselves on a bicycle etc. are asked to answer questions about the accident via an online questionnaire. The questionnaire has a map function where the accidents are located with coordinates.
It is of interest to investigate the spatial distribution of the accidents in order to understand more about critical points and imporant street characteristics which might increase or decrease the risk of accidents. The interesting thing is that the road networks is not continuous in space but lives on a metric graph.
Supporting courses
The most important supporting courses for working within computational statistics are TMA4300 Computational statistics and TMA4250 Spatial statistics, but other statistics courses are also useful.
Spring semester:
Autumn semester:
Previous Master students
- Sara Elise Wøllo (2022), "Correcting for under-reporting of violence against women in Italy using INLA".
- Julie Berg (2022), "Multistate models in survival analysis using INLA, applied on data for resuscitation after cardiac arrest".
- Helene Behrens (2021) "Bayesian Mortality Modeling with Linearized Integrated Nested Laplace Approximation".
- Marion Helen Røed (2021), "Spatial Extreme Value Modelling of Sea Level Data from the Oslo Fjord"
- August Sørli Mathisen (2020), " Inference on extreme hourly precipitation in Norway using INLA"
- Martin Outzen Berild (2020), "Integrated Nested Laplace Approximations within Monte Carlo Methods"
- Sigrid Leithe (2019) “Statistical Methods for the Analysis of Data with a Lower Limit of Detection”
- Anne Siri Fardal (2019), "A Bayesian Model for Prediction of Heat Consumption"
- Johan Øvstebø Birketvedt (2019), "Interval Censored Regular Vines with Application to Event-Based
Modelling of Precipitation and Temperature"