Improved sammon mapping method for visualization of multidimensional data

Authors:
Halina Kwasnicka;Pawel Siemionko
Affiliations:
Institute of Informatics, Wroclaw University of Technology, Wroclaw, Poland;Institute of Informatics, Wroclaw University of Technology, Wroclaw, Poland
Venue:
ICCCI'12 Proceedings of the 4th international conference on Computational Collective Intelligence: technologies and applications - Volume Part II
Year:
2012

Citing 3
Cited 0

Readings in information visualization: using vision to think

Readings in information visualization: using vision to think
Research Report: Volume Rendering for Relational Data

INFOVIS '97 Proceedings of the 1997 IEEE Symposium on Information Visualization (InfoVis '97)
A Nonlinear Mapping for Data Structure Analysis

IEEE Transactions on Computers

Quantified Score

Hi-index	0.00

Visualization

Abstract

Three improvements to the Sammon mapping method are proposed. Two of them concern calculation complexity reduction. Introducing the limit for delta parameter allows to eliminate error fluctuations during data projection. Calculating distances not for all data points but for the part of them results in important reduction of the calculation time without worsening the final results. The third improvement allows adding new data to the projected ones without recalculation of all data from the beginning. The paper presents details of the proposed improvements and the performed experimental study.