Decision Tree and Random Forest


#1

Can anyone explain and help me to understand why decision tree and random forest are Non-linear models ?


#2

A linear model is one which can be expressed using a linear equation (makes sense, right?).

On the other hand a decision tree cannot be expressed using an equation. DTree is more like a flow, based of the condition (ex.: Rain=Yes or No) the tree progresses until a leaf node is reached. This traversing might not be linear hence, non linear model.

Hope this helps.


#3

Thank you Nishant for the detailed explanation. It helped.


#4

One way to look at it is that a condition for a split in a decision tree can be something like “x>3”. This is not a linear condition since 3.0000000000001 is true but 3 is not…


#5

You can read these following links random-forest-and-gradient-boosting-machines
gradient-boosting-machines-example


#6

thank you Swapansh for the links