Scribd
Upload
35 views4 pages

ECE479579 Class Notes 03-14

The document discusses topics for an upcoming quiz, including AND/OR graphs and expert systems. It then covers imprecise reasoning in expert systems using confidence factors to represent the measure of belief in rule outputs. Several examples are provided to illustrate how confidence factors are calculated for rules connected by AND and OR operators.

Uploaded by

vincent xu
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
35 views4 pages

ECE479579 Class Notes 03-14

The document discusses topics for an upcoming quiz, including AND/OR graphs and expert systems. It then covers imprecise reasoning in expert systems using confidence factors to represent the measure of belief in rule outputs. Several examples are provided to illustrate how confidence factors are calculated for rules connected by AND and OR operators.

Uploaded by

vincent xu
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/ 4

03/14/2018 ECE479/579 - Lecture (21)

Quiz topics
1. AND /OR graph
2. Alpha Beta prune
3. Expert system (until lecture 20)

Understand
 what graph is
 what use for
 how to compute the cost

Imprecise Reasoning
 How to compute none probabilistically confidence level that certain searching using
expert system.
 For giving more background posting work Stanford university using the concept to write
one of the first expert system to help physicians prescribe particular type of antibiotic

Confidence Factor (CF) or Measure Of Believe (MB)


 Constrain 0 ≤ 𝑀𝐵 ≤ 1
Posing three questions

aggregate
1. What is the measure of believe (MB) of the input to a rule?

MB1
i1
i2 Output
MB2
i3
R
MB3
Having rule uses
AND logical conjunction
Input

What is the aggregate measure of believe?


Formal MB ( i1 & i2 & ……& in )= Min { MB1 , MB2………. MBn }

 Taking the aggregate to be the minimum the measure of confidence in the input
 Why taking the Min!

1
0.95 MB 0.45 MB

 R1 if X drives a Volvo AND X reads the New York Times (NYT) then X will vote Democrat

What is the aggregate input?


o Min {0.95 , 0.45} = 0.45 Why taking the Min!

Question 2 is
2. 1.a if having to compute the measure of believe of

MB ( i1 OR i2 OR ……OR in )= Max { MB1 , MB2………. MBn }

0.6 MB
 R1 if X does not like real estate developers

OR
0.7 MB

If X likes National Public Radio (NPR) then X will vote Democrat

i 0.7
R Output
Input
OR
 Max {0.7 , 0.6} = 0.6

 What is the measure of believe MB in the rule itself?


Taking
0.225 The confidence in the
confidence
input
R Output
0.45 * 0.5 = 0.225
output is half of the
which is confidence of input
0.5
0.45

Intuition factor
Output = Min (MB) * intuition factor

2
03/14/2018 ECE479/579 - Lecture (21)

Question 3 is
How to combine this?
 Multiple facts supporting the same conclusion what is my measure of believe in that
conclusion?

0.95 MB Volvo Vote


Min {MB} 0.45 * 0.5 = 0.225
0.45 MB NYT AND R1 D
E
M 0.442
0.5 O
Intuition factor
0.6 MB C
⌐Real Estate developer
Max {MB} 0.7 * 0.4 = 0.28
R
0.7 MB NPR OR R2 A
T
0.4
Intuition factor

Output = Min (MB) * intuition factor

Output = Max (MB) * intuition factor

What is measure of believe MB (Democrat, R1 & R2)


Referring to system is called MYCIN
MB (H, R1&R2) = MB (R1) + MB (R2) * (1 – MB (R1))

Hypothesis 0.225 + 0.28 * (1 – 0.225) = 0.442

 Different rules or independent rules supporting same conclusion.

3
Introduction to predicate logic
 Logic as representation method
o What is representation know
o How to describe UML!
 Structured representation method because it is lean object class and physical
entities and it has hierarchy class, subclass and decomposition.
Logic is called unstructured representation
o What do we know about logic?
o Vocabulary
 Each time using capital letters A , B ,….K  label for object

 Lower case x logic variable Logic world Real world


A-------------- -----------A
B-------------- -----------B
<X, A> means X  A x-------------- ---------
X A
True
o Predicates On(A, B) OR
False
 Predicate is expression whose value is true or false
 Dealing with binary logic not fuzzy logic

o Function y=On(x)
To express the relationship between object

o Basic operators
 & , ^ AND A ^ B Conjunction

 V OR AVB Disjunction

 ⌐ , ~ negation ⌐A It is true or false depends on A

 ══> implication A ═> B = ⌐A v B


equal

You might also like