Decision tree

yani_ca031yani_ca031 Member Posts: 16 Maven
edited November 2018 in Help

Hello everyone:

The problem that brings me here today should not be very complicated, but I can not think of any way to solve it.

The question is this: When we use a decision tree it returns a series of information that in my case it would be useful to know. First, it would be useful to know what is the way you are following to make a new classification; and second, I would like to know what is the distribution in that rule and the number of examples of the total that classify me in it. This is possible to see when we place the cursor on the final leaf of the tree on the diagram, but how could I know all this automatically if I want to classify a new example?

Answers

  • MartinLiebigMartinLiebig Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 3,503 RM Data Scientist

    Hi @yani_ca031,

    have a look at the Decision Tree to Example Set operator which is part of Converters Extension. This should return all the information you want.


    ~Martin

    - Sr. Director Data Solutions, Altair RapidMiner -
    Dortmund, Germany
  • yani_ca031yani_ca031 Member Posts: 16 Maven

    Thanks, Martin.

     

    You're right, but he does not return the information to me the way I need it. When I want to classify new examples I can not find a way to get all this automatically. I would have to extract this information from the rapidminer and then find where the current conditions coincide with the historical ones so that I can use the information. In other words, I would have to use "get decision tree path" to compare and know exactly where I am.

  • yani_ca031yani_ca031 Member Posts: 16 Maven

    Then I would have to extract the results and match them in excel for example.

  • MartinLiebigMartinLiebig Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 3,503 RM Data Scientist

    Hi,

    why do you want to switch platforms? Just read in the excel to RM, score it and write it.

     

    Best,

    Martin

    - Sr. Director Data Solutions, Altair RapidMiner -
    Dortmund, Germany
  • yani_ca031yani_ca031 Member Posts: 16 Maven

    Hi again Martin:

     

    I do not understand what you want to say. How i can?

     

    I need all this information in a single row. that is, I need to know what rule applies to the classification of the new example, the distribution of said rule, the number of examples that compose it and what proportion is of the total of examples of learning. My question is: Can I only do this if I extract this information separately and take it to excel to link all this on my own? 

  • yani_ca031yani_ca031 Member Posts: 16 Maven

    However, I can not find a way to do this without taking it elsewhere to compare it, find the coincidences and unite everything.

  • yani_ca031yani_ca031 Member Posts: 16 Maven

    Any idea?

Sign In or Register to comment.