Bayesian Inference for Spatio-Temporal Models
Shubin, Mikhail (2016)
Shubin, Mikhail
Helsingin yliopisto
2016
Julkaisun pysyvä osoite on
https://urn.fi/URN:ISBN:978-951-51-2290-2
https://urn.fi/URN:ISBN:978-951-51-2290-2
Tiivistelmä
The dissertation presents five problem-driven research articles, representing three research domains related to micro-organisms causing infectious disease. Articles I and II are devoted to the A(H1N1)pdm09 influenza (`swine flu') epidemic in Finland 2009-2011. Articles III and IV present software tools for analysing experimental data produced by Biolog phenotype microarrays. Article V studies a mismatch distribution as a summary statistic for the inference about evolutionary dynamics and demographic processes in bacterial populations.
All addressed problems share the following two features: (1) they concern a dynamical process developing in time and space; (2) the observations of the process are partial and imprecise.
The problems are generally approached using Bayesian Statistics as a formal methodology for learning by confronting hypothesis to evidence. Bayesian Statistics relies on modelling: constructing a generative algorithm mimicking the object, process or phenomenon of interest.
All addressed problems share the following two features: (1) they concern a dynamical process developing in time and space; (2) the observations of the process are partial and imprecise.
The problems are generally approached using Bayesian Statistics as a formal methodology for learning by confronting hypothesis to evidence. Bayesian Statistics relies on modelling: constructing a generative algorithm mimicking the object, process or phenomenon of interest.
Kokoelmat
- Kirjat [4204]