Aarhus University Seal / Aarhus Universitets segl

Causal Graphs

Britta Anker Bak
Lectures for students by students
Friday, 17 May, 2013, at 14:30-15:30, in Aud. D3 (1531-215)
"Correlation does not imply causation" is one of the golden sentences in statistics. Though true in the sense that correlation do not mean that one of two correlated events have to cause the other, it is not possible for two events to be correlated without the one causing the other or having a common cause.

In this talk we explore causal structures illustrated by Directed Acyclic Graphs (DAG). I will look upon to which extend causal structures can be determined from data. Furthermore it will be explained how knowing the causal structure will make it possible to determine the causal effect without randomized experiments, which are often unfeasible or even unethical in practice.

An introductury course in probability should be sufficient to follow the talk (eg. Introduktion til matematisk modellering).
Contact person: Thomas Lundsgaard Schmidt