BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY: Discovering Outcomes via Propensity Score Analysis of Social Medi
 a Timelines
DTSTART:20151210T110000
DTEND:20151210T120000
DTSTAMP:20260407T111037Z
UID:b687e0c398c170256311537d4b79dd53dbe3a0902c8a615bf7993db3
CATEGORIES:Conferences - Seminars
DESCRIPTION:Emre Kiciman\, Microsoft Research Redmond\nPeople make decisio
 ns every day\; some are significant and life-changing decisions and other
 s minor and mundane. Relying on their personal experience and knowledge\, 
 individuals rightly or wrongly choose what to do based on expectations of 
 their decision’s outcomes.  What if we could tell everyone what happene
 d when others were in the same situation?  With 10-100s of millions of pe
 ople regularly reporting the details of their real-world experiences in so
 cial media\, we believe there is an opportunity to mine the likely outcome
 s of situations and actions to help people make more informed decisions.\n
 We present a framework for extracting such outcomes from personal experien
 ces reported in social media timelines.  Applying stratified propensity s
 core analysis\, we isolate the outcomes of specific events from observed c
 onfounding factors.  We benchmark the quality and coverage of results ext
 racted from 3 months of Twitter data for 39 situations in a wide-range of 
 domains including social\, health and business topics.  We also demonstra
 te the application of this analysis technique in deeper case studies\, inc
 luding a study of the precursors of shifts towards suicide ideation among 
 pseudonymous posters in Reddit support communities.
LOCATION:BC 01 https://plan.epfl.ch/?room==BC%2001
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR