WEBVTT

00:00:00.930 --> 00:00:04.050
Welcome to the first lecture
of this class.

00:00:04.050 --> 00:00:06.860
You may be used to having a
first lecture devoted to

00:00:06.860 --> 00:00:09.670
general comments and motivating
examples.

00:00:09.670 --> 00:00:11.020
This one will be different.

00:00:11.020 --> 00:00:14.500
We will dive into the heart
of the subject right away.

00:00:14.500 --> 00:00:16.920
In fact, today we will
accomplish a lot.

00:00:16.920 --> 00:00:20.030
By the end of this lecture, you
will know about all of the

00:00:20.030 --> 00:00:23.120
elements of a probabilistic
model.

00:00:23.120 --> 00:00:26.060
A probabilistic model is a
quantitative description of a

00:00:26.060 --> 00:00:29.080
situation, a phenomenon,
or an experiment

00:00:29.080 --> 00:00:31.320
whose outcome is uncertain.

00:00:31.320 --> 00:00:34.850
Putting together such a model
involves two key steps.

00:00:34.850 --> 00:00:38.140
First, we need to describe the
possible outcomes of the

00:00:38.140 --> 00:00:39.320
experiment.

00:00:39.320 --> 00:00:42.910
This is done by specifying
a so-called sample space.

00:00:42.910 --> 00:00:46.170
And then, we specify a
probability law, which assigns

00:00:46.170 --> 00:00:50.060
probabilities to outcomes or
to collections of outcomes.

00:00:50.060 --> 00:00:53.360
The probability law tells us,
for example, whether one

00:00:53.360 --> 00:00:57.110
outcome is much more likely
than some other outcome.

00:00:57.110 --> 00:01:00.640
Probabilities have to satisfy
certain basic properties in

00:01:00.640 --> 00:01:02.070
order to be meaningful.

00:01:02.070 --> 00:01:05.069
These are the axioms of
probability theory.

00:01:05.069 --> 00:01:08.370
For example probabilities
cannot be negative.

00:01:08.370 --> 00:01:11.940
Interestingly, there will be
very few axioms, but they are

00:01:11.940 --> 00:01:14.890
powerful, and we will see
that they have lots of

00:01:14.890 --> 00:01:16.220
consequences.

00:01:16.220 --> 00:01:19.400
We will see that they imply
many other properties that

00:01:19.400 --> 00:01:22.480
were not part of the axioms.

00:01:22.480 --> 00:01:25.510
We will then go through a couple
of very simple examples

00:01:25.510 --> 00:01:27.789
involving models with
either discrete

00:01:27.789 --> 00:01:29.730
or continuous outcomes.

00:01:29.730 --> 00:01:32.690
As you will be seeing many times
in this class, discrete

00:01:32.690 --> 00:01:35.070
models are conceptually
much easier.

00:01:35.070 --> 00:01:39.160
Continuous models involve some
more sophisticated concepts,

00:01:39.160 --> 00:01:43.270
and we will point out some of
the subtle issues that arise.

00:01:43.270 --> 00:01:45.880
And finally, we will talk a
little bit about the big

00:01:45.880 --> 00:01:49.560
picture, about the role of
probability theory, and its

00:01:49.560 --> 00:01:51.220
relation with the real world.