University of Konstanz
Algorithmics
Prof. Dr. Ulrik Brandes

Seminar "Multidimensional Scaling"

+++ News +++

Apr 18: Seminar in D 210

Multidimensional Scaling (MDS) is a family of methods for analysis and graphical representation of high-dimensional data. Using dimensionality reduction, similarity data is represented as distances among points in a geometric space. Originating from behavioral sciences, it nowadays widespread in many other areas of appliation: statistics, marketing, sociology, etc.; particularly, it is suited for many tasks in information visualization. This seminar covers some of the major developments in MDS, including models, algorithms, and examples.

Dates

Seminar (U. Brandes, C. Pich) Thursday, 10.15 – 11.45 (D 210)

Organization

This seminar will cover the following topics based on original research articles and monographs:

Each talk is of maximum length 60 minutes; after the presentation all seminar participants will discuss content and presentation technique.

All participants are required to compile a written extended abstract/paper about the topic within four weeks after the presentation containing main results, examples, and references to original work. We recommend to use (pdf)LaTeX for document preparation; formulas, figures, and tables should not be scanned/copied but cleanly and carefully reproduced if possible.

Talks

Talks will be assigned in the first meeting.

Date Speaker Topic
2007-04-19 · First meeting and talk assignment
2007-05-23 Christian Pich Applications of MDS to Graph Drawing
2007-06-14 Simon Endele Prokrustes-Analyse
2007-07-12 Lars Volkhardt Modellierung und Analyse asymmetrischer Daten

General Reading

Further Links