Write a Python program that returns a classification given
by the CART decision tree algorithm. The input and output
files and format is the same as assignments 1, 2, and 3.

This assignment is worth 5% of the total grade. We will test
it on a simple example to determine if your program gives the
correct output. Test your program with a simple XOR example.

Submit your assignment by copying it into the directory
/afs/cad/courses/ccs/f15/cs/675/001/<ucid>.
For example if your ucid is abc12 then copy your Perl script into 
/afs/cad/courses/ccs/f15/cs/675/001/abc12.

Your completed script is due before 1pm on Dec 9th 2015. 
Submit a hardcopy in class. 

High level pseudocode:

Recurse(data d, node n):

    (1) For each column j:
	        
	(1) Find the value that gives the best gini split of 
	the data d into a partition of two sets

	(2) To evaluate the gini of a split use the formula

	gini = (lsize/rows)*(lp/lsize)*(1 - lp/lsize) + (rsize/rows)*(rp/rsize)*(1 - rp/rsize);

	where lsize is the size of the left partition, lp is the 
	proportion of -1 labels in the left partition, rsize is the 
	size of the right partition, rp is the proportion of -1 
	labels in the right partition, and rows is the total number of
	datapoints in the dataset d (passed to the function)

    (2) Let column k give the best split. 
    
    (3) If the gini of the best split < 0.05 then stop and return
	Else :

	(1) Create child nodes l and r. Let dl be the datapoints that 
	are on the left side of the split and dr be the ones on the right side.

	(2) Recurse(dl, l)

	(3) Recurse(dr, r)