1
00:00:00,287 --> 00:00:04,350
[MUSIC]

2
00:00:04,350 --> 00:00:06,710
We've now learned
decision trees from data.

3
00:00:06,710 --> 00:00:09,300
Let's look at how we can
make predictions from it.

4
00:00:09,300 --> 00:00:14,010
So given the model that we've learned from
the data, we're going to predict y hat.

5
00:00:14,010 --> 00:00:17,530
As we discussed in the beginning of
the module, it's a pretty simple process

6
00:00:17,530 --> 00:00:21,050
of traversing down the tree in
order to make that prediction.

7
00:00:21,050 --> 00:00:25,410
Given that particular input xi, so going
back to that first example where Credit

8
00:00:25,410 --> 00:00:29,420
was poor, Income was high, and the Term
was 5 years, we go down the branch of

9
00:00:29,420 --> 00:00:34,090
the decision tree that is associated
with those particular inputs, and

10
00:00:34,090 --> 00:00:38,670
we make a prediction that y hat i is Safe.

11
00:00:38,670 --> 00:00:40,630
This would be explicit,

12
00:00:40,630 --> 00:00:45,240
even that prediction algorithm is
what's called a recursive algorithm.

13
00:00:45,240 --> 00:00:47,780
I'm trying to predict, for
a particular decision tree,

14
00:00:47,780 --> 00:00:52,010
I start with the top node and
my input set.

15
00:00:52,010 --> 00:00:54,810
And then I say,
if the current node is a leaf, return,

16
00:00:54,810 --> 00:00:57,040
predict whatever is on that leaf.

17
00:00:57,040 --> 00:01:01,150
If it's not a leaf, then pick the child
that agrees with the input and

18
00:01:01,150 --> 00:01:02,810
then recurse on that child.

19
00:01:02,810 --> 00:01:06,140
So return the prediction the child makes.

20
00:01:06,140 --> 00:01:10,675
And then that process is going to
unroll that path down the tree in order

21
00:01:10,675 --> 00:01:12,357
to make the prediction.

22
00:01:12,357 --> 00:01:16,739
[MUSIC]