SURVEY ON MULTIDIMENSIONAL VISUAL ANALYSIS TECHNIQUES

PhD Qualifying Examination


Title: "SURVEY ON MULTIDIMENSIONAL VISUAL ANALYSIS TECHNIQUES"

by

Mr. Nan CAO


Abstract:

Multidimensional data are commonly used to represent both structured and 
unstructured information. Unfortunately, multidimensional data analysis is 
very challenging as the data are usually complex in nature, huge in 
amount, and contain both statistical and geometrical features. Information 
visualization can be of great value for multidimensional data analysis as 
it can represent multidimensional data in intuitive ways and also support 
explorative visual analysis which keeps humans in the loop. During the 
past decades, various visualization designs and related analysis 
techniques have been proposed for multidimensional data.

In this article, we review multidimensional visualizations and related 
explorative analysis techniques over past 40 years. We first introduce the 
history of this field and traditional taxonomies. Then we present a new 
taxonomy which classifies the visualizations into two categories: item 
packing and statistical embedding. Item packing techniques focus on 
displaying every single attribute of the data while statistical embedding 
methods mainly represent the multivariate statistical features. Major 
visualization techniques in each category are reviewed in detail. After 
that, explorative visual analysis systems for analysis tasks such as 
dimension reduction, visual clustering analysis and visual diagnostics are 
discussed. Finally, we conclude with several potential research directions 
and topics in this field.


Date:                   Monday, 26 September 2011

Time:                   2:00pm - 4:00pm

Venue:                  Room 3501
                         lifts 25/26

Committee Members:	Dr. Huamin Qu (Supervisor)
                         Prof. Chi-Keung Tang (Chairperson)
 			Dr. Pedro Sander
 			Prof. Qiang Yang


**** ALL are Welcome ****