Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

Decision Trees

I'm new to data mining and just trying to figure out how it works. I've started with Decision Trees for binomial classification.

One thing that I have found that has surprised me is that the tree can change quite a bit as I add to my data set. Even if I only add examples that will all predict negative, the tree still changes. My overall sense is that the tree becomes more arbitrary, even after pruning, but that might only be my current project. People around me who see the variability of the tree structure are immediately full of doubts about the decisions that are rendered. I can't say I blame them.

Can someone explain what is going on, in layman's terms?

Find more posts tagged with

AI Studio

Accepted answers

All comments

fras

You may treat decision trees as "Data Driven Documents". Yes, they are very sensitive to data and your findings are not a surprise. Pruning and
defining a max depth of the tree could help. To check how well your data is represented or just (over) fitted you may also apply an X-Validation operator.
Happy mining !

Legacy User

Hello to everyone,

I need little bit help in decision tree. I was importing csv to rapidminer and get a tree for the Titanic project, but there is missing something. Can someone help me please?

Thank you!

LBWood

I'm just a beginner myself. I taught myself using the book Data Mining for The Masses. Can you describe your steps and where you get the problem?