Case and the decision tree algorithm

Discussion in 'microsoft.public.sqlserver.datamining' started by Joe, Nov 10, 2005.

  1. Joe

    Joe Guest

    When it is needed to use case and nested table in the decision tree algorithm?.
    Is only useful when you need to relate tables but is not the same concept
    that in Association rules, where case is the entity that we want to analyze.
    (i.e. the Basket in the market basket analysis)

    Thanks in advance.
    Joe.
     
    Joe, Nov 10, 2005
    #1
    1. Advertisements

  2. The idea of the nested table is to group together a set of attributes from
    the same domain that are associated with a case. This concept is applicable
    across all algorithms.

    For example, you might have demographic information about customers like
    age, marital status, income as well as their favorite TV shows and you want
    to predict their gender. This would be a good candidate for a decision tree
    model with "Favorite TV Shows" as a nested table. If you didn't have nested
    tables, you would need to have a boolean case-level column for each show -
    "Likes Friends", "Likes Oprah", "Likes Smallville" and so forth.
     
    Raman Iyer [MS], Nov 10, 2005
    #2
    1. Advertisements

  3.  
    anonymous_user, Nov 18, 2005
    #3
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.