Statistics Seminar: Sisi Ma
Ford Hall, Room 115
Title: Multiple Markov Boundaries: the Good, the Bad, and the Ugly
The Markov boundary is a minimal set of variable conditioned on which all the remaining variables in the data set are rendered statistically independent of the response variable T. For certain distributions, only one single Markov boundary is present for a given response variable, however, it is possible that multiple variable sets are Markov boundaries. I will introduce (multiple) Markov boundaries theory. I will also discuss the implications of the presence of multiple Markov boundaries for predictive and causal analytics.