The Interval-Censored Biplot

ISSN： 1061-8600

Source： Journal of Computational and Graphical Statistics, Vol.22, Iss.1, 2013-01, pp. : 123-134

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

The principal components biplot is a useful visualization tool for the exploration of a samples by variables data matrix. In several data analysis situations, the data values are interval censored so that only the interval of a data value is available, but not the value itself. For such data, we propose the interval-censored biplot (IC-Biplot), a new exploratory and graphical method that is an extension of the principal component analysis biplot. It provides not only a two-dimensional graphic representation of respondents and their attributes, but also point estimates for the data values that are constrained to be in their interval. Two applications of the IC-Biplot are discussed. The first application considers data on emergence times of permanent teeth focusing on the pattern of emergence. The IC-Biplot confirms rank orders suggested earlier in the literature. Goodness-of-fit measures show that the present model seems to fit these data very well. The second application discusses a regular sample by attribute matrix from the literature on characteristics of several types of oils. This article has online supplementary materials.