large set or small set for clustering

  • Loren Engrav

    Loren Engrav - 2010-02-25

    this may be trivial to you all but
    We have a system of about 24000 probes.  Filters reduce that to about 2000.  Statistical q reduces that to about 200.
    If we wish to explore hierarchical and k-means clustering, is it most "proper" to use the 2000 set or the 200 set?

  • raktim

    raktim - 2010-02-25

    Sorry, this is difficult to comment on as everything depends on your study and how and what you are asking.
    I would recommend you to talk a bioinformatician for help.

    Thanks for using MeV


Log in to post a comment.