WEBVTT

00:00:00.320 --> 00:00:04.019
In this last lecture of this
unit, we continue with some of

00:00:04.019 --> 00:00:07.670
our earlier themes, and then
introduce one new notion, the

00:00:07.670 --> 00:00:11.480
notion of independence
of random variables.

00:00:11.480 --> 00:00:14.820
We will start by elaborating a
bit more on the subject of

00:00:14.820 --> 00:00:17.890
conditional probability
mass functions.

00:00:17.890 --> 00:00:21.070
We have already discussed the
case where we condition a

00:00:21.070 --> 00:00:23.710
random variable on an event.

00:00:23.710 --> 00:00:27.690
Here we will talk about
conditioning a random variable

00:00:27.690 --> 00:00:31.510
on another random variable,
and we will develop yet

00:00:31.510 --> 00:00:34.420
another version of the total
probability and total

00:00:34.420 --> 00:00:36.810
expectation theorems.

00:00:36.810 --> 00:00:40.920
There are no new concepts
here, just new notation.

00:00:40.920 --> 00:00:45.070
I should say, however, that
notation is important, because

00:00:45.070 --> 00:00:48.590
it guides you on how to think
about problems in the most

00:00:48.590 --> 00:00:51.410
economical way.

00:00:51.410 --> 00:00:55.190
The one new concept that we will
introduce is the notion

00:00:55.190 --> 00:00:57.960
of independence of
random variables.

00:00:57.960 --> 00:01:01.200
It is actually not an entirely
new concept.

00:01:01.200 --> 00:01:04.580
It is defined more or less the
same way as independence of

00:01:04.580 --> 00:01:09.110
events, and has a similar
intuitive interpretation.

00:01:09.110 --> 00:01:13.130
Two random variables are
independent if information

00:01:13.130 --> 00:01:16.960
about the value of one of them
does not change your model or

00:01:16.960 --> 00:01:20.090
beliefs about the other.

00:01:20.090 --> 00:01:23.130
On the mathematical side, we
will see that independence

00:01:23.130 --> 00:01:25.740
leads to some additional
nice properties

00:01:25.740 --> 00:01:27.010
of means and variances.

00:01:29.560 --> 00:01:32.640
We will conclude this lecture
and this unit on discrete

00:01:32.640 --> 00:01:36.190
random variables by considering
a rather difficult

00:01:36.190 --> 00:01:38.750
problem, the hat problem.

00:01:38.750 --> 00:01:41.910
We will see that by being
systematic and using some of

00:01:41.910 --> 00:01:45.170
the tricks that we have learned,
we can calculate the

00:01:45.170 --> 00:01:48.910
mean and variance of a rather
complicated random variable.