Cluster analysis and related data visualization tools are vital components in efforts to interpret multidimensional or multifield scientific and biomedical data. An initial selection of these tools are newly part of Civilized Software's MLAB product for DOS-based personal computers. MLAB is a mathematical modeling language, originally developed at NIH for use on DEC-10 computers. The original DEC-10 MLAB, which included a set of cluster analysis functions (""""""""CLAB""""""""), has been widely used by medical researchers, scientists, and statisticians, as testified to by over 500 citations in the literature. It is expected that PC MLAB will be even more widely utilized, both by existing users as DEC-10s are phased out, and by new users. PC MLAB has been significantly enhanced in overall functionality and usability beyond the DEC-10 version, and contains an important initial subset of CLAB functions, including principle component analysis, K-means clustering, and Fisher ranking.
The aim of this proposal is to program and integrate the remaining DEC-10 CLAB algorithms into PC MLAB: methods for creating and visualizing minimum spanning trees and five different types of dendrograms. This will be followed during Phase II by extensions and new operators that will provide a broad base of cluster analysis functionality.
