The analysis of high-dimensional data is an important, yet inherently difficult problem. Projection techniques
such as PCA, MDS, and SOM can be used to map high-dimensional data to 2D display space. However,
projections typically incur a loss in information. Therefore, uncertainty usually exists when evaluating the
precision of the projection against the original high-dimensional data characteristics. While the output quality
of these techniques can be discussed in terms of algorithmic assessment, visualization is often helpful for better
understanding the results.
In this paper, we address the assessment of projection precision by an approach integrating an appropriately
designed projection precision measure directly into the projection visualization. To this end, a
exible projection
precision measure is de ned that allows the user to balance the degree of locality at which the measure is evalu-
ated. Several visual mappings are designed for integrating the precision measure into the projection visualization
at various levels of abstraction. The techniques are implemented in a fully interactive system which is practically
applied on several data sets. We demonstrate the usefulness of the approach for visual analysis of classi ed and
clustered high-dimensional data sets. We thereby show how our novel interactive precision quality visualization
system helps to examine preservation of closeness of the data in original space into the low-dimensional space.
| 