what is the meaning of "number of prepruning alternatives" of decision tree parameters?

lg21clg21c Member Posts: 1 Contributor I
edited January 2 in Help
It is explain in the Help--"When split is prevented by prepruning at a certain node this parameter will adjust the number of alternative nodes tested for splitting. Occurs as prepruning runs parallel to the tree generation process. This may prevent splitting at certain nodes, when splitting at that node does not add to the discriminative power of the entire tree. In such a case, alternative nodes are tried for splitting."


but I can not unerstand,Please give me a more clearer explanation
Tagged:

Answers

  • Telcontar120Telcontar120 Moderator, RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,045   Unicorn
    Here is my attempt to summarize based on my understanding of how this is implemented in RapidMiner.
    The DT algorithm proceeds at each node by looking for the variable and the split value that provides the best improvement in the growth criterion (e.g., accuracy, or information gain ratio or whatever you have selected). Pre-pruning restricts the tree from growing based on certain constraints you enter such as the number of minimum number of cases in each leaf.   So it could be that the best option in terms of the growth criterion violates one of the pre-pruning constraints, in which case it won't actually grow that branch if you have activated pre-pruning. 
     Rather than stopping prematurely in such a scenario, the "number of pre-pruning alternatives" allows the algorithm to examine other potential variables and/or splits that might allow the tree to continue growing because those alternatives, although they don't necessarily improve the growth criterion as much, do not violate one of the other pre-pruning constraints. 
    Brian T.
    Lindon Ventures 
    Data Science Consulting from Certified RapidMiner Experts
    sgenzer
Sign In or Register to comment.