Master of Science Eirik Myrvoll-Nilsen will Friday May 22 at 12.15 publically defend his thesis for the PhD degree in Science.
Title of the thesis:
«Efficient Bayesian analysis of long memory processes applied to climate»
Popular scientific abstract:
Since the industrial revolution the temperature on Arth has increased substantially due to increased radiative forcing. This is defined as the imbalance between incoming and outgoing radiation in the climatic system and is affected by factors such as amission of climate gases. Moreover, climatic time series are known to exhibit long memory, which implies that even distant variables can be correlated. This can make obtaining statistical inference very computationally difficult.
We overcome this by approximating long memory processes as a mixture of short memory processes. This was found to be remarkable accurate, even when using just four short memory processes. This grants us a major reduction in computational cost and allows us to incorporate this into a general modeling framework which can be used to estimate how susceptible the climate system is to increased CO2-emission, perform temperature prediction given possible scenarios for future forcing and analyse local time series data.
Supervisors:
Evaluation Committee:
Opposition ex auditorio:
If you have any questions for the candidate during the public defence, please send an e-mail to: trygve.johnsen@uit.no. He will announce them to the candidate.
Leader of the public defence:
Professor Trygve Johnsen, Department of Mathematics and Statistics, Faculty of Science and Technology, UiT
Trial lecture:
The trial lecture is held Friday May 22 at 10.15 and will be streamed. The title of the trial lecture:
«An introduction to Gaussian Markov random fields»