AIFA: Fuzzy Relation
07/03/2024
Koustav Rudra
Relations
• AI = {Akash, Ashish}
• MATH = {Sundar, Sekhar}
• AI x MATH = {(Akash,Sundar), (Akash,Sekhar), (Ashish,Sundar), (Ashish, Sekhar)}
• Relation R = Close Friends
• 𝑅 ⊆ { 𝐴𝑘𝑎𝑠ℎ, 𝑆𝑢𝑛𝑑𝑎𝑟 , (𝐴𝑠ℎ𝑖𝑠ℎ, 𝑆𝑒𝑘ℎ𝑎𝑟)}
• How to go beyond absolute membership?
Fuzzy Relations
• A Fuzzy relation for N sets is defined as an extension of the crisp relation to include membership
grade
• 𝑅 = {𝜇! (𝑥" , … , 𝑥# )/(𝑥" , … , 𝑥# )|𝑥$ ∈ 𝑋$ }
Fuzzy Relation
• Fuzzy relation describes interactions between variables
• It defines the mapping of variables from one fuzzy set to another
• Like crisp relation, we can also define the relation over fuzzy sets
• Main fuzzy operations and compositions are followings:
• Union
• Intersection
• Complement
• Difference
• Max-min composition
• Max-product composition
Operations on Fuzzy Set
• Union:
• 𝐴∪𝐵
• 𝜇%∪' 𝑥 = max(𝜇% 𝑥 , 𝜇' 𝑥 ), ∀𝑥 ∈ 𝑋 [Every member of A and B]
• Example:
• The First Fuzzy Set is : {'a': 0.2, 'b': 0.3, 'c': 0.6, 'd': 0.6}
• The Second Fuzzy Set is : {'a': 0.9, 'b': 0.9, 'c': 0.4, 'd': 0.5}
• Fuzzy Set Union is : {'a': 0.9, 'b': 0.9, 'c': 0.6, 'd': 0.6}
• Intersection:
• 𝐴∩𝐵
• 𝜇%∪' 𝑥 = m𝑖𝑛(𝜇% 𝑥 , 𝜇' 𝑥 ), ∀𝑥 ∈ 𝑋 [Every member of A and B]
• Example:
• The First Fuzzy Set is : {'a': 0.2, 'b': 0.3, 'c': 0.6, 'd': 0.6}
• The Second Fuzzy Set is : {'a': 0.9, 'b': 0.9, 'c': 0.4, 'd': 0.5}
• Fuzzy Set Intersection is : {'a': 0.2, 'b': 0.3, 'c': 0.4, 'd': 0.5}
Operations on Fuzzy Set
• Complement
• 𝐴(
• 𝜇%! 𝑥 = 1 − 𝜇% (𝑥)
• Example:
• The Fuzzy Set is : {'a': 0.2, 'b': 0.3, 'c': 0.6, 'd': 0.6}
• Fuzzy Set Complement is : {'a': 0.8, 'b': 0.7, 'c': 0.4, 'd': 0.4}
• Difference:
• 𝐴−𝐵
• 𝜇%)' 𝑥 = m𝑖𝑛(𝜇% 𝑥 , 1 − 𝜇' 𝑥 ), ∀𝑥 ∈ 𝑋 [Every member of A and B]
• Example:
• The First Fuzzy Set is : {'a': 0.2, 'b': 0.3, 'c': 0.6, 'd': 0.6}
• The Second Fuzzy Set is : {'a': 0.9, 'b': 0.9, 'c': 0.4, 'd': 0.5}
• Fuzzy Set Intersection is : {'a': 0.1, 'b': 0.1, 'c': 0.6, 'd': 0.5}
Operations on Fuzzy Set
• Containment
• 𝐴⊆𝐵
• 𝜇% 𝑥 ≤ 𝜇' 𝑥 ), ∀𝑥 ∈ 𝑋
• Equality:
• 𝐴=𝐵
• 𝜇% 𝑥 = 𝜇' 𝑥 , ∀𝑥 ∈ 𝑋
Compositions of two relations
• Max-Min Composition
• Given two relation matrices R and S, the max-min composition is defined as 𝑇 = 𝑅 ° 𝑆
• 𝑇 𝑥, 𝑧 = max{min{𝑅 𝑥, 𝑦 , 𝑆 𝑦, 𝑧 𝑎𝑛𝑑 ∀* ∈ 𝑌}}
Max-Min Composition
1 3 5 1 3 5
• 𝑋 = 1,3,5 ; 𝑌 = {1,3,5} • 𝑅= 1 0 1 0 • 𝑆= 1 0 1 1
• 𝑅= 𝑥, 𝑦 𝑦 = 𝑥 + 2 ; 𝑆 = {(𝑥, 𝑦)|𝑥 < 𝑦} 3 0 0 1 3 0 0 1
• 𝑅 𝑎𝑛𝑑 𝑆 𝑖𝑠 𝑜𝑛 𝑋 × 𝑌 5 0 0 0 5 0 0 0
• 𝑅 = { 1,3 , (3,5)}
• 𝑆 = { 1,3 , 1,5 , (3,5)}
1 3 5
1 0 0 1
𝑅°𝑆
3 0 0 0
5 0 0 0
Fuzzy Cartesian Product
• A is a fuzzy set on the universe of discourse X with 𝜇% 𝑥 |𝑥 ∈ 𝑋
• B is a fuzzy set on the universe of discourse Y with 𝜇& 𝑦 |𝑦 ∈ 𝑌
• 𝑅 = 𝐴×𝐵 ⊂ 𝑋×𝑌
• 𝜇' 𝑋, 𝑌 = 𝜇%×& 𝑋, 𝑌 = min 𝜇% 𝑥 , 𝜇& 𝑦
• 𝐴 = { 𝑎), 0.2 , 𝑎*, 0.7 , (𝑎+, 0.4)}, B = { 𝑏), 0.5 , 𝑏*, 0.6 }
𝑏! 𝑏"
𝑎! 0.2 0.2
𝑅 = 𝐴×𝐵
𝑎" 0.5 0.6
𝑎# 0.4 0.4
Operations on Fuzzy Relations: Max-Min
𝑦! 𝑦"
𝑧! 𝑧" 𝑧#
𝑥! 0.5 0.1
𝑦! 0.6 0.4 0.7
𝑥" 0.2 0.9
𝑦" 0.5 0.8 0.9
𝑥# 0.8 0.6
𝑧! 𝑧" 𝑧#
𝑥! 0.5 0.4 0.5
𝑥" 0.5 0.8 0.9
𝑥# 0.6 0.6 0.7
Compositions of two relations
• Max-Product Composition
• Given two relation matrices R and S, the max-min composition is defined as 𝑇 = 𝑅 ° 𝑆
• 𝑇 𝑥, 𝑧 = max{{𝑅 𝑥, 𝑦 ∗ 𝑆 𝑦, 𝑧 𝑎𝑛𝑑 ∀* ∈ 𝑌}}
Operations on Fuzzy Relations: Max-Product
𝑦! 𝑦" 𝑧! 𝑧" 𝑧# 𝑧! 𝑧" 𝑧#
𝑥! 0.6 0.3 𝑦! 1 0.5 0.3 𝑥! 0.6 0.3 0.21
𝑥" 0.2 0.9 𝑦" 0.8 0.4 0.7 𝑥" 0.72 0.36 0.63
Fuzzy Rule
07/03/2024
Koustav Rudra
Fuzzy Rule
• A Fuzzy implication of the form:
• If X is A then Y is B
• A and B are two linguistic variables defined by Fuzzy sets A and B
• On the universe of discourses X and Y, respectively
• Antecedent: X is A
• Consequence: Y is B
If-Then Rules
• Use Fuzzy sets and Fuzzy operators as the subjects and verbs of fuzzy logic to form rules
• If x is A then Y is B
• Where A and B are linguistic terms defined by fuzzy sets on the sets X and Y respectively
• If velocity is small, then current needed is small
• If temp is high, put cooler as moderate
Fuzzy Implication: Example 1
• If pressure is high then temperature is low
• If mango is yellow then mango is sweet else mango is sour
• The Fuzzy Implication is denoted as 𝑅: 𝐴 → 𝐵
If-Then Rules
• If x is A is true then …
• If 70 is high is true then …
0.5
0.7
If-Then Rules
• If x is A is and y is B then …
• If 70 is high and 5 is medium then …
0.7 0.8
0.7
Evaluation of Fuzzy Rules
• In Boolean logic: p ⟹ 𝑞
• If p is true then q is true
• In Fuzzy logic: p ⟹ 𝑞
• If p is true to some degree then q is true to some degree
• 0.5p ⟹ 0.5𝑞
• Partial premise implies partially
• How?
Max-min rule of composition
• Given N observations 𝐸$ over X and hypothesis 𝐻$ over Y we have N rules:
• If 𝐸" 𝑡ℎ𝑒𝑛 𝐻"
• If 𝐸+ 𝑡ℎ𝑒𝑛 𝐻+
• If 𝐸# 𝑡ℎ𝑒𝑛 𝐻#
• 𝜇, = max{min 𝜇-" , min 𝜇-# , … , min(𝜇-$ )}
Fuzzy Cartesian Product
• A is a fuzzy set on the universe of discourse X with 𝜇% 𝑥 |𝑥 ∈ 𝑋
• B is a fuzzy set on the universe of discourse Y with 𝜇& 𝑦 |𝑦 ∈ 𝑌
• 𝑅 = 𝐴×𝐵 ⊂ 𝑋×𝑌
• 𝜇' 𝑋, 𝑌 = 𝜇%×& 𝑋, 𝑌 = min 𝜇% 𝑥 , 𝜇& 𝑦
• 𝐴 = { 𝑎), 0.2 , 𝑎*, 0.7 , (𝑎+, 0.4)}, B = { 𝑏), 0.5 , 𝑏*, 0.6 }
𝑏! 𝑏"
𝑎! 0.2 0.2
𝑅 = 𝐴×𝐵
𝑎" 0.5 0.6
𝑎# 0.4 0.4
Fuzzy Implication: Example 2
• Suppose P and T represent Pressure and Temperature
• P = {1,2,3,4}
• T = {10, 15, 20, 25, 30, 35, 40, 45, 50}
• 𝑇,./, = { 20,0.2 , 25,0.4 , 30,0.6 , 35,0.6 , 40,0.7 , 45,0.8 , (50,0.8)}
• 𝑃012 = { 1,0.8 , 2,0.8 , 3,0.6 , (4,0.4)}
• if temperature is HIGH, then pressure is LOW, 𝑅: 𝑇,./, → 𝑃012
1 2 3 4
20 0.2 0.2 0.2 0.2
25 0.4 0.4 0.4 0.4
30 0.6 0.6 0.6 0.4
35 0.6 0.6 0.6 0.4
40 0.7 0.7 0.6 0.4
45 0.8 0.8 0.6 0.4
50 0.8 0.8 0.6 0.4
Set of Support
07/03/2024
Koustav Rudra
Set of Support
• It is a Crisp set
• Consist of elements whose membership values in the corresponding
Fuzzy Set is greater than zero
𝛂 Cut / Horizontal Cut
• It is a crisp set
• Consist of elements whose membership values in the corresponding Fuzzy
Set is greater than 𝛂
1
A
𝜇 $%&& (ℎ)
0
A𝛂
Thank You
