Decision Tree

sudheendrasudheendra Member Posts: 22 Maven
in Help
Hai,

I have an unstructured Text document.I want to represent the related words as a Decision Tree, ie  If I use ChangeAttributeRole and input a particular word as label can I obtain a Decision Tree on the basis of this document?

Thanks,
Sekhar

Answers

  • haddockhaddock Member Posts: 849 Maven
    Hi there,
    I want to represent the related words as a Decision Tree
    I'm having difficulty understanding this, could you give us some form of example?
  • sudheendrasudheendra Member Posts: 22 Maven
    Hai,

    I am attaching some of my contents.
    Suppose it contains "He suffered head trauma and is still experiencing vision problems as a result of this accident" which is  repeating many  times in my Text  Data. So I need to get the relation between "accident", "vision" and "headtrauma" or in other sense if my decision variable is "vision" then the tree should contain "accident" and other related attributes.

    Thanks,
    Sekhar

  • haddockhaddock Member Posts: 849 Maven
    So the root node of the decision tree would be what?
  • sudheendrasudheendra Member Posts: 22 Maven
    Hai,

    I have only one document. I am extracting ideas from the document.I dont know whether it is possible to take the entire document as root.In the above example is it possible to take "accident "as root? For the first time I am applying Decison Tree in Text document.So your help will be highly useful  :)

    Thanks,
    Sekhar
  • landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    I doubt, that decision trees will be of great help for your problem. I don't see there any classification problem. Please remember, a classification problem is about assigning a class label, for example "red" or "green". What should be this label in your case?

    Greetings,
      Sebastian
  • sudheendrasudheendra Member Posts: 22 Maven
    Hi Sebastain,

    I have a 50 text files. I need to classify the total documents into ''Good' and "Bad" on the basis of words ( ie If the documents contain loss,injury,damage etc then I need to classify it into "Bad", if the documet contain issued,payment,closed etc then I need to classify it into "Good")
    How can I proceed further?

    Thanks,
    Sekhar
  • haddockhaddock Member Posts: 849 Maven
    Have you read this?

    http://kent.dl.sourceforge.net/project/yale/2.%20Text%20Plugin/4.6/rapidminer-text-4.6-tutorial.pdf
Sign In or Register to comment.