Build a decision tree (DT) classfier using the following three strategies for stopping growth

of a tree: i) threshold on the decrease in impurity ii) threshold on the number of data

vectors at a node iii) grow the tree as much as possible and then prune.

Experiment with at least two different impurity functions.

handle some errors in the data like absurd values and/or missing data for

some attribute values. Carefully, report how you have handled such data errors and what

is the reason for the choice you made.

In all cases report five fold cross validated values for accuracy.