0
$\begingroup$

In most of the popular packages (rpart etc.), the decision tree split is based on information gain, entropy or Gini index.

Is there a popular decision tree package ( either 'R' or python) where the node split quality can be specified using chi-squared statistic?

$\endgroup$
2
$\begingroup$

Base on another post about significance test for stopping decision tree in stats stackexchange website. CHAID, QUEST, GUIDE are three options. The QUEST was provide in binary format, but not available now.

The usage for an implementation of CHAID in R is expained here. To install it on R

install.packages("partykit")
install.packages("CHAID", repos="http://R-Forge.R-project.org")
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.