WEBVTT
00:00:00.600 --> 00:00:01.940
Now let's do an example.
00:00:01.940 --> 00:00:06.740
So here's your reminder of the formula and suppose this is my data.
00:00:06.740 --> 00:00:11.040
I have four data points and for each data point I have three features,
00:00:11.040 --> 00:00:17.570
the grade of the train, the bumpiness, whether there's a speed limit,
00:00:17.570 --> 00:00:20.709
and then the speed that the car goes.
00:00:21.950 --> 00:00:23.440
And i'm just making up some data here.
00:00:23.440 --> 00:00:26.890
So let's start out at the top of the decision tree here.
00:00:26.890 --> 00:00:31.960
So I have two slow examples and two fast examples.
00:00:31.960 --> 00:00:36.200
Slow, slow, fast, fast.
00:00:36.200 --> 00:00:39.199
So the first question is, what's the entropy of this node?
00:00:40.470 --> 00:00:42.590
So let's do this piece by piece.
00:00:42.590 --> 00:00:46.880
How many of the examples in this node are slow?
00:00:46.880 --> 00:00:47.810
Write your answer here.