Unlike classical data which are points in p-dimensional space, symbolic data are hypercubes or Cartesian products of distributions in p-dimensional space. For example, the observations could be interval-valued. While some symbolic data sets arise naturally, many more emerge today after aggregation of larger data sets. Applications abound. We look at some aspects unique to symbolic data that require special attention in any subsequent analysis. Then, we outline approaches involved in principal component analysis of interval data, and clustering for distributional data.

This is a joint event with SSAI.


6:00pm - 6:30pm: Refreshments (@ Staff Common room, 3rd floor, The Red Centre)

6:30pm - 7:30pm: Lecture

7:45pm onwards: Dinner (at a nearby restaurant)


Professor Lynne Billard

Research Area

University of Georgia, USA


Wed, 18/05/2016 - 6:00pm


RC-4082, The Red Centre, UNSW