Special Interest Group

Techniques in Data Matching and Contextual Search

ACS & DAMA Joint event

National Press Club, 16 National Circuit, Barton, ACT, 2601
Tue 19 Sep 2017 05:00 PM AEST
Duration: 2.5 hours
Register by Tue 19 Sep 2017 02:00 PM AEST
In Person
CPD Hours: 2
Skills Level: Information governance (IRMG) -> Level 4

About this event

Registration from 5pm, Presentation at 5:30, Networking canapes and drinks from 6:30.

Data is never as clean and accurate as we might wish. That's not always because of data management errors. The data may have been recorded accurately from an unreliable source, or represent one of the several ways to record the same facts (variable spelling, abbreviations or alternate place names or address formats), or have an intrinsic uncertainty, e.g. the estimated age of a person.
When we want to search, or have more than one data source that we must harmonise, we cannot expect that all values will match exactly. We must draw on a range of techniques for approximate matching, and may need the additional validation of each match by the use of contextual data.

This presentation covers the main techniques used in handling uncertainty in search and data matching. Techniques from large-scale batch data matching cover pre-processing, indexing, comparison and classification. Searching also needs to rank results from approximate matching, partial matching, and from a contextual search that spans both direct and indirectly related data. Data relationships must be conformed around the main subjects of inquiry; Person, Location, Case, etc.

The techniques are those investigated in Factil's Feasibility Study for the BRII national child safety data sharing project. 

Speakers

Speaker
Clifford Heath
Proudly sponsored by

Event Location

National Press Club, 16 National Circuit, Barton, ACT, 2601
Read