Why are DBI using Local Random Seed and Determined Good Start Value same in K-Means

anindyaranianindyarani Member Posts: 3 Newbie
Hi, i'm working on text clustering using K-Means and Singular Value Decomposition (SVD). And i'm using parameter Local Random Seed and Determined Good Start Value to show the different. But the DBI value generated using Local Random Seed is always the same as Determine Good Start Value even though I have tried to enter a random number. Can anyone explain this?

Answers

  • MartinLiebigMartinLiebig Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 3,314 RM Data Scientist
    Hi,
    can you provide an example process here?
    I think the reason can be two fold. Either a bug, or simply that the algorithm converges nicely to exactly the same means no matter where you start. I mean, thats what you ideally want anyhow.

    BR,
    Martin
    - Head of Data Science Services at RapidMiner -
    Dortmund, Germany
  • anindyaranianindyarani Member Posts: 3 Newbie
    edited January 23
    the design process and the setting parameter (k-Means


    parameter SVD


    so i'm doing an algorithm comparison between K-Means and K-Means &SVD, I'm also doing variations on the Determine Good Start Value and Local Random Seed sections. So later I will compare a total of 4 algorithms with different parameter settings.
Sign In or Register to comment.