certainty factors - .certainty theory certainty factor (cf ), value to measure degree of belief from
Post on 13-Mar-2019
216 views
Embed Size (px)
TRANSCRIPT
CERTAINTY FACTORS
Lecture #6
Expert System Lab Work
Uncertanity
There is a uncertainty in the mindOf experts when make a decisionaccordance their expertise
Representation Uncertainty
Basis Concept of ProbabilityCertainty FactorBayesian ReasoningFuzzyNeural NetworkGenetic Algorithm
Causes of Uncertainty
Not completely and uncertainty information Unknown dataUnify the different of view point of expertsImprecise language (always,often,seldom,some
times)
5
Causes Uncertainty
Impreciselanguage
Term Always Very often Usually Often Generally Frequently Rather often About as often as not Now and then Sometimes Occasionally Once in a while Not often Usually not Seldom Hardly ever Very seldom Rarely Almost never Never
Mean value 99 88 85 78 78 73 65 50 20 20 20 15 13 10 10
7 6 5 3 0
Term Always Very often Usually Often
Generally Frequently Rather often
About as often as not Now and then Sometimes Occasionally Once in a while Not often Usually not Seldom Hardly ever Very seldom Rarely Almost never Never
Mean value 100 87 79 74 74 72 72 50 34 29 28 22 16 16 9 8 7 5 2 0
Milton Hakel (1968) Ray Simpson (1944)
Certainty Theory
Certainty factor (cf ), value to measure degree of belief from expert.
Maximum value of cf is +1.0 (definitely true) and minimum -1.0 (definitely false).
Two aspects: Certainty in the EvidenceThe evidence can have a certainty factor attached
Certainty in Rule
1. Model Net Belief (Shortliffe dan Buchanan)
CF (Rule)=MB(H,E)-MD(H,E)
MB(H,E) =
M D(H,E) =
Computation of Certainty Factors
1 ; if P(H)=1
max[P(H|E),P(H)]-P(H)/1-P(H) ; Otherwise
1 ; if P(H)=0
min[P(H|E),P(H)]-P(H)/1-P(H) ; Otherwise
Certainty Factors
Uncertain Term Certainty FactorDefinitely not -1.0Almost certainly not -0.8Probably not -0.6Maybe not -0.4Unknown -0.2 to +0.2Maybe +0.4Probably +0.6Almost certainly +0.8Definitely +1.0
2. Direct Interview with Expert
Example :
Expert :If headache and have a cold and fever, then most possibility is influenza
Rule : IF evidence 1= headache AND evidence 2= have a cold AND evidence 3= fever
THEN seek=influenza(cf = 0.8)
Expert System With CF
In Expert System with CF, knowledge base composed of set of rules with syntax:
IF THEN {cf }
CF refers to degree of belief for hypothesis H when evidence E occured.
Degree of belief for hypothesis H when evidence E occured.cf (H,E) = cf (E) * cf(Rule)
e.g:IF sky is clearTHEN the forecast is sunny {cf 0.8}
With cf of sky is clear is 0.5cf (H,E) = 0.5 * 0.8 = 0.4
Example
Multiple Antecedents
How the CF if we have two evidence?With conjunction (i.e. and)
Use minimum cf of evidenceWith disjunction(i.e. or)
Use maximum cf of evidence
Conjunctive Antecedents - Example
Conjunctive Rules: cf(H, E1 and E2 and Ei) = min{cf(E1 , E2 , Ei )}
*cf(Rule)
IF there are dark clouds E1AND the wind is stronger E2THEN it will rain {cf 0.8}
If cf(E1) = 0.5 and cf(E2) = 0.9, then
cf(H, E) = min{0.5, 0.9} * 0.8 = 0.4
Disjunctive Antecedents - Example
Disjunctive Rules:cf(H, E1 or E2 or Ei) = max{cf(E1 , E2 , Ei )}
*cf(Rule)
IF there are dark clouds E1OR the wind is stronger E2THEN it will rain {cf 0.8}
If cf(E1) = 0.5 and cf(E2) = 0.9, then
cf(H, E) = max{0.5, 0.9} * 0.8 = 0.72
How to determine CF if two rules have similar conclusion?
Similarly Concluded Rules
Rule 1: IF weatherperson predicts IF weatherperson predicts rain (Erain (E11) THEN it will rain) THEN it will rain{{cfcf 11 0.7} 0.7}
Rule 2:IF farmer predicts rain IF farmer predicts rain (E(E22) THEN it will rain) THEN it will rain{{cfcf 22 0.9} 0.9}
Similarly Concluded Rules - Example
Two rules with similar conclusion:
IF weatherperson predicts IF weatherperson predicts rain Erain E11THEN it will rainTHEN it will rain{{cfcf 11 0.7} 0.7}
IF farmer predicts rain IF farmer predicts rain E E22THEN it will rainTHEN it will rain{{cfcf 22 0.9} 0.9}
Suppose cf(ESuppose cf(E11) = 1.0 and cf(E) = 1.0 and cf(E2 2 ) = 1.0) = 1.0
cfcf 1(H1, E1) = cfcf(E1) * cfcf(Rule1)
cfcf 1(H1, E1) = 1.0 * 0.7 = 0.7
cfcf 2 (H2, E2) = cfcf(E2) * cfcf(Rule2)
cfcf 2(H2, E2) = 1.0 * 0.9 = 0.9
Similarly Concluded Rules
Formula :
Similarly Concluded Rules
New CF value from facts above above can be formulated:cf (cf1,cf2) = cf1+cf2*(1-cf1)
= 0.7+0.9(1-0.7)= 0.97
EXERCISE
Determine New Value of CF from some facts in below :1. IF fever
THEN typhus {cf -0.39}2. IF amount of thrombosis is low
THEN typhus {cf -0.51}3. IF the body is weak THEN typhus {cf 0.87}4. IF diarhea/constipation THEN typhus {cf 0.63}
Combining CF with CLIPS
Combining CF with CLIPS
Combining CF with CLIPS
Combining CF with CLIPS
Combining CF with CLIPS
Lakukan Beberapa Langkah
- Load -> Reset -> (agenda)- Perhatikan urutan rule yang dijalankan
Rule dengan Salience yang lebih besar akan dieksekusi terlebih dahulu
Rule dengan Salience yang lebih besar akan dieksekusi terlebih dahulu
Modifikasi Salience dari Rule
(defrule both-positif(declare (salience 6) .... .... ....
(defrule both-negatif(declare (salience 10) .... .... ....
Lakukan langkah yang sama : Load -> Reset -> Agenda
Lihat pada Agenda:
Akan terlihat bahwa rule dengan salience tinggi, akan dijalankan terlebih dahulu
Combining CF with CLIPS
Cf = 0.8390766
Slide 1Slide 2Slide 3Slide 4Slide 5Slide 6Slide 7Slide 8Slide 9Slide 10Slide 11Slide 12Slide 13Slide 14Slide 15Slide 16Slide 17Slide 18Slide 19Slide 20Slide 21Slide 22Slide 23Slide 24Slide 25Slide 26Slide 27Slide 28