Nacionalni inštitut za biologijo Vas vabi, da se udeležite dveh predavanj:

Lecture 1: Bayesian Thinking


Lecture 2: Bayesian Evaluation of Informative Hypotheses

ki ju bo predstavil

prof. dr. Herbert Hoijtink

Utrecht University, The Nederlands

Predavanji bosta v četrtek, 18. oktobra 2012 z začetkom ob 13. uri
v predavalnici B1 v Biološkem središču, Večna pot 111, Ljubljana.
(drugo predavanje ob 14. uri, vmes bo odmor)


Lecture 1: Bayesian Thinking (ob 13:15)

Using vignettes instead of formulas two aspects in which Bayesian statistics differs from classical inferential approaches will be highlighted. First of all, the difference between frequentist and Bayesian probabiliy will be elaborated. Secondly, the difference between null-hypothesis significance testing and model selection will be discussed. This lecture is concluded with an example in which Bayesian probabilities are used to select the best of two hypotheses that are not the traditional null or alternative hypotheses, but hypotheses reflecting a researchers expectations with respect to the population of interest.

Lecture 2: Bayesian Evaluation of Informative Hypotheses (ob 14:15)

In this lecture it will be argued that researchers are almost never interested in the traditional null and alternative hypotheses. Instead they are interested in so-called informative hypotheses, that is, hypotheses that are representations of the expectations researchers have with respect to the relations between the variables in their population of interest. Examples of informative hypotheses will be given. Furthermore, the Bayes factor will be introduced in a non-technical manner, and it will be shown how informative hypotheses can be evaluated using the Bayes factor.

PS: prvo predavanje je bolj splošno in provokativno, drugo pa malo bolj poglobljeno.

Andrej Blejec
National Institute of Biology
Vecna pot 111 POB 141, SI-1000 Ljubljana, SLOVENIA

e-mail: andrej.blejec@nib.si
URL: http://ablejec.nib.si
tel: + 386 (0)59 232 789
fax: + 386 1 241 29 80
Organizer of
Applied Statistics 2012 conference