Tuesday, June 21st, 2011, 1:17 am
Ramble About GMDS and GPCA (Ongoing Research)
Filed under:
VER the past few months I have been running many experiments and compiling about 200 pages of text and graphs to document this. Today I made a short explanatory video that was unplanned and totally unscripted (I tend to prefer spontaneity, I loathe staging or reading from outlines/talking points/scripts).
This video is a bit of a test intended to get the hang of the recording software. These are some preliminary thoughts on the approach taken by GMDS (generalised multidimensional scaling) and the one adopted by GPCA (generalised principal component analysis) proponents. Admittedly, I am only at the early stages of properly learning about both. However, the basic principles as they are applied to distances (in space, not within data instances embedded in the shown hyperspace) are analogous at some underlying level, i.e. they measure something which is theoretically a surrogate of one another. The premises are hinged upon unification or use of one method to complement another, perhaps just comparing the results of each one in isolation.
