Binning of Kernel-based projection pursuit indices in XGobi

Sigbert Klinke, Dianne Cook

Research output: Contribution to journalArticleResearchpeer-review

5 Citations (Scopus)

Abstract

The software XGobi offers a selection of different index functions to be used in exploratory projection pursuit (EPP). EPP is a method for finding interesting low-dimensional (in this case two-dimensional) projections of highdimensional data. We discuss the inclusion of two additional index functions which make use of several standard techniques to improve the speed of the indices based on kernel density estimation. Unfortunately, the speed improvements do not apply to derivative calculations, but then are seldomly required in XGobi so speed does not matter much.

Original languageEnglish
Pages (from-to)363-369
Number of pages7
JournalComputational Statistics and Data Analysis
Volume25
Issue number3
DOIs
Publication statusPublished - 21 Aug 1997
Externally publishedYes

Keywords

  • Exploratory projection pursuit
  • Fast implementation
  • Highdimensional data analysis
  • Kernel density estimation

Cite this