1
$\begingroup$

I am applying the AD Tree algorithm & this is the tree visualization of the output:

enter image description here

I can't understand the values in the decision nodes (-0.4,0.541,-0.882...), How are these calculated? & how did we calculate the root node's score?

Are predicate conditions (<127.5..) formed by entropy splitting mechanism?

This is an image of the output:

enter image description here

Any help is appreciated, cannot find any AD Tree output analysis document!!

$\endgroup$
1
$\begingroup$

As per the original paper here on Alternative Tree it says,

"We give a different representation of the same classification rule.... The decision node is identical to what we had before , while the prediction node is associated with a "real valued number"...... The classification that is associated with the path is NOT the label of the leaf, instead that is the sign of the sum of the predictions along the path.for example, the classification of the instance a=b=0.5 is sign(0.5-0.7-0.2)=sign(-0.4)=-1..."

Now there is optimization of it too here which is available here and is implemented by WEKA and JBoost . And of course there is always -Wiki

enter image description here

$\endgroup$

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.