BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bayesian modeling of networks in complex business intelligence pro
 blems - Daniele Durante (Università degli Studi di Padova )
DTSTART:20160825T131000Z
DTEND:20160825T135000Z
UID:TALK67057@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:<span>Co-authors: Sally Paganin (University of Padova\, Dept. 
  of Statistical Sciences)\, Bruno Scarpa (University of Padova\, Dept. of 
  Statistical Sciences)\, David B. Dunson (Duke University\, Dept. of  Stat
 istical Science) <br></span> <br>Complex network data problems are increas
 ingly common in many fields of  application. Our motivation is drawn from 
 strategic marketing studies monitoring  customer choices of specific produ
 cts\, along with co-subscription networks  encoding multiple purchasing be
 havior. Data are available for several agencies  within the same insurance
  company\, and our goal is to efficiently exploit  co-subscription network
 s to inform targeted advertising of cross-sell strategies  to currently mo
 no-product customers. We address this goal by developing a  Bayesian hiera
 rchical model\, which clusters agencies according to common  mono-product 
 customer choices and co-subscription networks. Within each cluster\,  we e
 fficiently model customer behavior via a cluster-dependent mixture of late
 nt  eigenmodels. This formulation provides key information on mono-product
  customer  choices and multiple purchasing behavior within each cluster\, 
 informing targeted  cross-sell strategies. We develop simple algorithms fo
 r tractable inference\, and  assess performance in simulations and an appl
 ication to business  intelligence <br><br>Related Links <ul> <li><a target
 ="_blank" rel="nofollow">http://arxiv.org/abs/1510.00646</a>  - Arxiv vers
 ion of the manuscript&nbsp\;</li></ul>
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR