BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:A scaling approach to record linkage - Harvey Goldstein (Universit
 y of Bristol\; University College London)
DTSTART:20160913T140000Z
DTEND:20160913T143000Z
UID:TALK67333@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:<span>Co-authors: Mario Cortina-Borja (UCL)\, Katie Harron  (L
 SHTM) <br></span> <span><br>With increasing availability of large data set
 s derived from administrative  and other sources\, there is an increasing 
 demand for the successful linking of  these to provide rich sources of dat
 a for further analysis. The very large size  of such datasets and the vari
 ation in the quality of the identifiers used to  carry out linkage means t
 hat existing approaches based upon &lsquo\;probabilistic&rsquo\;  models c
 an make heavy computational demands. They are also based upon  questionabl
 e assumptions. In this paper we suggest a new approach\, based upon a  sca
 ling algorithm\, that is computationally fast\, requires only moderate amo
 unts  of storage and has intuitive appeal. A comparison with existing meth
 ods is  given.&nbsp\;</span>
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR