Tuesday, May 19, 2009

Multidimension Scaling

Multidimension Scaling is a technique for projecting multi-dimensional points in a 2-d plan, so that the distances as preserved as much as possible.

Here you have a C++ code skeleton with STL and boost.


  1. So in your code you're mapping 2D-points to 2D-points?

  2. Aren't you creating an array of 100 2D-points (test.cpp) and then mapping them on to another set of 2D-points?

  3. Hello!

    I have tried your MDS, it works well. I am no programmer (occasional scripts/batch), but I have a few questions;
    -how could I use/change it for a CSV input data (table with 1000+ rows, 5+ columns)?
    -possible integration with CUDA?

    Thank you again!

  4. This is not exactly a Multidimension Scaling. this routine mimic the functioning of Multidimension Scaling in case where the input dimension is 2 and the output dimension is also 2. this skeleton code cannot be extended to support two different dimension (the input and output must always be of same dimension)