BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20250916T143205EDT-3402gn78zv@132.216.98.100
DTSTAMP:20250916T183205Z
DESCRIPTION:Leila Golparvar\, PhD\n\nPostdoctoral Researcher\, Department o
 f Mathematics and Statistics\, ŗŚĮĻ²»“ņģČ University\n\nCausal Structure Learn
 ing and Propensity Score Adjustment\n\nJoint work with: D.A. Stephens and 
 R. Platt\n\nALL ARE WELCOME\n\nAbstract:\n\nMathematical representation of
  causal dependencies among a set of variables via graphs has been in the s
 tatistical literature for more than a century. A graph G = (V \, E) consis
 ts of a vertex set V = {X1\,...\,Xp} representing observed variables\, and
  an edge set E representing structural links between the variables. Predic
 tĀ­ing the eļ¬€ect of manipulations from non-experimental data therefore ofte
 n involves two steps: ļ¬rst\, discovery of causal structures\, represented 
 by directed acyclic graphs (DAGs)\, and second identiļ¬cation and estimatio
 n of causal parameters.\n\nIn this talk\, we will adopt a frequentist cons
 traint-based approach to discover the causal graph\, and use the PC algori
 thm for causal discovery. The PC algorithm is based on sequential testing 
 of partial correlations between variables and a search strategy that ident
 iļ¬es the presence and absence of structural links. To explore the use of t
 he PC algorithm in selection of confounders\, we conduct a Monte Carlo sim
 ulation study. It is known that adjusting for all confounders will elimina
 te bias\, but additionally adjusting for predictors of outcome that are un
 related to treatment will lead to estimates with lower variance. Results s
 how that PC algorithm works very well when the sample size is moderate to 
 large.\n\nKEYWORDS: Causality\; Confounding\; Counterfactual\; Directed ac
 yclic graph\; PC-algorithm.\n\nĀ \n\nĀ \n\nĀ \n
DTSTART:20160126T203000Z
DTEND:20160126T213000Z
LOCATION:Room 24\, Purvis Hall\, CA\, QC\, Montreal\, H3A 1A2\, 1020 avenue
  des Pins Ouest
SUMMARY:Biostatistics Seminar
URL:/epi-biostat-occh/channels/event/biostatistics-sem
 inar-258036
END:VEVENT
END:VCALENDAR