Classification week 4 overfitting in decision trees

gistfile1.txt

Learning simpler decision trees

early stopping -> limit depth, use classification error on limit depth of tree, if very few data points

Pruning: simplify tree after learning algorithm terminates
cost function

step 1: consider a split
step 2: computer total cost C(t) of split
c(t) = Error(t) + lamda * Number of leaf(t) 
prune if total cost is lower than total cost of tree

in real world, data is really messy

3 strategies skip, imput, adapt learning algorithm to be robust to missing values

feature split selection