Thanks to visit codestin.com
Credit goes to www.scribd.com

Scribd
Upload
88 views1 page

Tutorial 1: Bayes Theorem and Bayesian Inference: I L R R L R R

This document introduces a Bayesian network for classifying images as containing cats or not based on eye appearance, eye separation, eye difference, and fur color variables. It provides the structure of the Bayesian network and examples of the prior and conditional probabilities that define the network. It also asks the reader to quantify one of the variables, calculate the missing prior and conditional probabilities from example data, and then use the network to calculate the probability of a cat being present given values for some of the variables in a new image.

Uploaded by

Peter Holc
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
88 views1 page

Tutorial 1: Bayes Theorem and Bayesian Inference: I L R R L R R

This document introduces a Bayesian network for classifying images as containing cats or not based on eye appearance, eye separation, eye difference, and fur color variables. It provides the structure of the Bayesian network and examples of the prior and conditional probabilities that define the network. It also asks the reader to quantify one of the variables, calculate the missing prior and conditional probabilities from example data, and then use the network to calculate the probability of a cat being present given values for some of the variables in a new image.

Uploaded by

Peter Holc
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

Tutorial 1: Bayes theorem and Bayesian inference

In lecture 2 we introduced a Bayesian network for recognising cats in pictures:


Variable Interpretation Type Value
C Cat Discrete (2 states) True or False
E Eyes Discrete (3 states) Probable, Possible, Unlikely
S Eye separation Continuous S = (S
i
2 (R
l
+ R
r
))/R
r
D Eye difference Continuous |(R
l
R
r
)/R
r
|
F Fur colour Discrete (20 states) Histogram of pixel hues
We then choose a method of quantisation of the variable S into seven states as follows:
[below 1.5][1.5 0.75][0.75 0.25][0.25 0.25][0.25 0.75][0.75 1.5][above1.5]
1. Suggest a way of quantising the variable the variable D into four states.
2. Given the following data set, calculate the prior probability P(C) and the link matrices P(D|E) and
P(E|C) in the same format as that given for P(S|E) and P(F|C) (NB The states of D are labeled
d
1
, d
2
, d
3
, d
4
etc).
c
1
e
1
s
2
d
1
f
5
c
1
e
1
s
2
d
2
f
6
c
2
e
1
s
3
d
2
f
4
c
1
e
1
s
2
d
3
f
3
c
1
e
1
s
3
d
1
f
4
c
2
e
2
s
1
d
1
f
7
c
1
e
2
s
3
d
2
f
10
c
1
e
2
s
4
d
3
f
4
c
1
e
3
s
5
d
1
f
8
c
2
e
3
s
4
d
2
f
3
c
2
e
3
s
6
d
3
f
1
c
1
e
3
s
7
d
1
f
2
c
2
e
3
s
1
d
4
f
9
c
2
e
3
s
2
d
4
f
8
c
2
e
3
s
7
d
4
f
1
P(S|E) =

P(s
1
|e
1
) P(s
1
|e
2
) P(s
1
|e
3
)
P(s
2
|e
1
) P(s
2
|e
2
) P(s
2
|e
3
)
P(s
3
|e
1
) P(s
3
|e
2
) P(s
3
|e
3
)
P(s
4
|e
1
) P(s
4
|e
2
) P(s
4
|e
3
)
P(s
5
|e
1
) P(s
5
|e
2
) P(s
5
|e
3
)
P(s
6
|e
1
) P(s
6
|e
2
) P(s
6
|e
3
)
P(s
7
|e
1
) P(s
7
|e
2
) P(s
7
|e
3
)

0 0.33 0.14
0.6 0 0.14
0.4 0.34 0
0 0.33 0.14
0 0 0.14
0 0 0.14
0 0 0.28

P(F|C) =

P(f
1
|c
1
) P(f
1
|c
2
)
P(f
2
|c
1
) P(f
2
|c
2
)
P(f
3
|c
1
) P(f
3
|c
2
)
P(f
4
|c
1
) P(f
4
|c
2
)
P(f
5
|c
1
) P(f
5
|c
2
)
P(f
6
|c
1
) P(f
6
|c
2
)
P(f
7
|c
1
) P(f
7
|c
2
)
P(f
8
|c
1
) P(f
8
|c
2
)
P(f
9
|c
1
) P(f
9
|c
2
)
P(f
10
|c
1
) P(f
10
|c
2
)

0 0.3
0.125 0
0.125 0.14
0.25 0.14
0.125 0
0.125 0
0 0.14
0.125 0.14
0 0.14
0.125 0

3. Given your answer to 2 and that for a particular picture S = s


4
, D = d
2
and F = f
3
calculate the
probability of there being a cat in the picture. (First calculate the likelihood of E).
DOC493: Intelligent Data Analysis and Probabilistic Inference Tutorial 1 1

You might also like