1
00:00:00,000 --> 00:00:06,835
For Homework six, you will need to
implement a programming assignment to do

2
00:00:06,835 --> 00:00:14,351
with Bayesian networks.
This is the Bayesian network that we'll be

3
00:00:14,351 --> 00:00:19,567
evaluating.
This is a simple diagnostic network where

4
00:00:19,567 --> 00:00:28,641
the variable are characteristics about the
patient such as whether they smoke or not,

5
00:00:28,641 --> 00:00:37,102
Whether they have visited Asia or not.
Some symptoms such as whether or not they

6
00:00:37,102 --> 00:00:44,403
have dyspnea, or whether or not they have
tested positive for an X-ray, which shows

7
00:00:44,672 --> 00:00:51,924
occlusion in the lung area, and there are
three possible conditions that they may

8
00:00:51,924 --> 00:00:56,669
have, which is tuberculosis, lung cancer,
or bronchitis.

9
00:00:56,669 --> 00:01:03,473
And our goal will be to write a program
that will evaluate the a posteriori

10
00:01:03,473 --> 00:01:10,720
probability of each of the three possible
diseases given some combination of

11
00:01:10,720 --> 00:01:14,364
evidence.
For example, you maybe given that the

12
00:01:14,364 --> 00:01:19,990
patient has visited Asia and has a
positive X-ray, and then you need to

13
00:01:19,990 --> 00:01:26,709
figure out which disease is most likely.
The probabilities associated with this

14
00:01:26,709 --> 00:01:30,990
network are given as follows.
They're also available in this paper,

15
00:01:30,990 --> 00:01:33,845
where this example was originally
published.

16
00:01:33,845 --> 00:01:40,003
Let me explain what these things mean.
Alpha is the variable about Asia which has

17
00:01:40,003 --> 00:01:46,693
the probability that somebody has actually
visited Asia is 0.01. As a consequence,

18
00:01:46,693 --> 00:01:53,268
the probability of not visiting Asia that
is the probability of not A is 0.99. So,

19
00:01:53,268 --> 00:01:57,437
those probabilities are implied in this
description.

20
00:01:57,437 --> 00:02:03,210
So, this is not the, all the possible
values that you will need to encode.

21
00:02:03,210 --> 00:02:08,823
Please note that. Similarly, the
probability of tuberculosis given that

22
00:02:08,823 --> 00:02:14,275
somebody's visited Asia, is five%..
Whereas, in general, if they have not

23
00:02:14,275 --> 00:02:19,593
visited Asia, it's one%..
As a consequence of this, the probability

24
00:02:19,593 --> 00:02:24,779
that somebody does not have tuberculosis
if they have visited Asia is 95%.

25
00:02:24,779 --> 00:02:30,819
And the probability that they do not have
tuberculosis if they have not visited Asia

26
00:02:30,819 --> 00:02:34,158
is 99%..
So, the converse probabilities need to be

27
00:02:34,158 --> 00:02:37,853
inferred for each possible case in this
description.

28
00:02:37,853 --> 00:02:43,488
Don't forget that in your encoding.
Once you've encoded this, you will be

29
00:02:43,488 --> 00:02:50,201
given, evidence and asked questions about
the a posteriori or rather the most likely

30
00:02:50,201 --> 00:02:54,728
explanation that is the best diagnosis
given the evidence.

31
00:02:54,728 --> 00:03:00,973
For example, given no evidence whatsoever,
what is the chance that you would have

32
00:03:00,973 --> 00:03:04,954
tuberculosis?
And the answer is just over one%, if you

33
00:03:04,954 --> 00:03:11,143
believe this belief network.
Your program can encode these

34
00:03:11,143 --> 00:03:19,273
probabilities in tables and use SQL, the
way of we described in class, to evaluate

35
00:03:19,273 --> 00:03:24,500
the required a posteriori probability for
each condition.

36
00:03:25,240 --> 00:03:32,577
Using SQL makes it easier.
And this network is certainly not too big

37
00:03:32,577 --> 00:03:39,325
for any SQL engine to handle.
Sql Lite3 in Python, which is built in, is

38
00:03:39,325 --> 00:03:44,944
a good thing to use, or you can use any
SQL engine, or you can encode the network

39
00:03:44,944 --> 00:03:50,631
directly using any algorithm that you may
wish to read up, but the one which we have

40
00:03:50,631 --> 00:03:54,875
covered in class is SQL.
So, this is the assignment.

41
00:03:54,875 --> 00:04:01,060
Please implement your program, try it out
for different possible conditions, and

42
00:04:01,060 --> 00:04:04,505
then open Homework six to try and answer
it.

43
00:04:04,505 --> 00:04:11,004
It will be a timed homework, as before, so
that I am sure you're actually using your

44
00:04:11,004 --> 00:04:14,918
program.
But there will be enough time to run your

45
00:04:14,918 --> 00:04:21,182
program, type in the conditions that are
given, and get the answer and answer the,

46
00:04:21,182 --> 00:04:24,120
the homework.
Good luck.