Universe of Discourse Membership Function: Adriano Joaquim de Oliveira Cruz - NCE e IM, UFRJ Adriano@

Summary
Fuzzy Sets - Introduction Classical Sets

Classical Sets and Membership Functions
Classical Sets Operations
Adriano Joaquim de Oliveira Cruz – NCE e IM, UFRJ Classical Set Properties
adriano@nce.ufrj.br Fuzzy Sets
Membership Functions and Fuzzy Sets
If you only have a hammer, everything is
a nail. Fuzzy Sets Terminology
@2001 Adriano Cruz NCE e IM - UFRJ Fuzzy Sets Introduction 2
Crips Sets Classical Sets

Universe of Discourse Membership Function
– It is the space where all set elements are – Defines whether a given element belongs
defined to a set
– For instance: 1 x ≥1.70
χ A (X) = 
• Human heights: 0 <= h <= 2.5m 0 x < 1.70
• room temperatures: -70o<=temp<=70o 1
0
1,70 altura
@2001 Adriano Cruz NCE e IM - UFRJ @2001 Adriano Cruz NCE e IM - UFRJ
Classical Sets/Problems Truth Tables

Classical sets have problems when applied to a A B | A∪ B A∩ B A
wide class of real problems. 0 0 | 0 0 1
The difficult problem of defining the frontier 0 1 | 1 0 1
between two sets (e.g. high/ not high) is called
1 0 | 1 0 0
Sorites Paradox, due to the dialectic Eubulides
from Miletus, opponent of Aristotle. 1 1 | 1 1 0
The paradox is stated as follows: 1 → true
“Is one grain of sand a heap? Are two grains of sand a 0 → false
heap? ...”
Which mathematical functions could be used to
represent the union, intersection and
complement?
1
Classical Sets Properties Classical Sets Properties
Involution A = A Idempotência A∪ A = A
Comutatividade A ∪ B = B ∪ A A∩ A = A
A∩ B = B ∩ A Identidade A∪∅ = A
Associatividade A ∪ ( B ∪ C ) = ( A ∪ B) ∪ C A∩∅ = ∅
A ∩ ( B ∩ C ) = ( A ∩ B) ∩ C A∪ X = X
Distributividade A ∪ (B ∩ C ) = ( A ∪ B) ∩ ( A ∪ C ) A∩ X = A
A ∩ (B ∪ C ) = ( A ∩ B) ∪ ( A ∩ C ) De Morgan A∩ B = A∪ B
A∪ B = A∩ B
The Laws of Aristotle The Laws of Aristotle
“Everything must be or not be, either in the The intersection of a set and its complement
present or in the future.” is empty.
The Union of a set to its complement yields
the Universe. A∩ A = ∅
A∪ A = X
This is called the law of non-
non-contradiction.
contradiction
This is called the law of the excluded middle.

middle
@2001 Adriano Cruz NCE e IM - UFRJ Fuzzy Sets Introduction 9 @2001 Adriano Cruz NCE e IM - UFRJ
Fuzzy Sets Fuzzy Sets - definition

The membership function of elements in a A fuzzy set can be represented by an
fuzzy set A is characterised by ordered set of pairs:
µ (.) : X → [0,1] Value
that maps every element from the set X into a
real number in the interval [0,1]. A = {( x, µ A ( x)) | x ∈ X }
1.0
Fuzzy Set Membership Universe of
0.5 function Discourse
0
1,70 1,80 altura
2
Membership degree Membership
Membership degree:
degree an element belongs to a One element may belong partially to more
given set by a degree of certainty. than one set
Some elements are more representative of – children={Pedro, Ana, Paulo, Marta}
the set main idea than others. – adolescents = {Pedro, Mateus, Joaquim}
– Excellent students={(Pedro,0.8), (Ana,0.9),
(Paulo,0.9), (Marta,1.0)} – children(Pedro)=0.2
– Very high = {(Oscar,0.95), (Michael – adolescents(Pedro)=0.8
Jordan, 0.95), (Junior Baiano,0.8)}
It isn’t probability! It isn’t probability!

Belonging to the set of high people with a A bottle full of an unknown liquid has 95%
membership degree of 0.25 indicates probability of being pure poison.
separation from the ideal definition of high Another full bottle of a very similar liquid
people by 0.75. belongs to the set of pure water to a degree
of 0.95, which means 5% poison.
A degree value of 0.25 does not mean that a This concentration of poison does not kill,
person this high may be found with however you will be very sick, may be a
probability 0.25 in the universe of people. doctor will be required.
If you know which bottle to choose, you
understood the differences.
Fuzzy Sets Representation 1 Fuzzy Sets Representation 2
Ordered Pairs:
Pairs A fuzzy set may represented Indicated as the union of all elements
by a set ordered pairs
The first element is the element itself and the
second its membership degree in the set.
A = ∑ µ A ( xi ) / xi
xi∈A
João = 1.65 m; Ana = 1.70 m; Oscar= 1.80 m
High = 0.25/João + 0.5/Ana + 1.0/Oscar
High = {(João, 0.25), (Ana,0.5), (Oscar,1.0)}
3
Fuzzy Sets Representation 3 Example of Membership Functions
Membership function:
function a fuzzy set may be
represented by a function that maps its elements
into real numbers in the interval [0,1].
µ(x)
1.0
0 altura ≤ 1.5
5 × altura − 7.5 Medium heights
 1.5 ≤ altura ≤ 1.7
µ media (altura) = 
− 5 × altura − 9.5 1.7 < altura < 1.9
0 altura ≥ 1.9
1.5 1.7 1.9 Heights
Unimodal Functions Singletons

A function is unimodal if
A Function is a singleton whenever
∀x1 , x2 ∈ X , ∀λ ∈ [0,1] : µ (λx1 + (1 − λ ) x2 ) ≥ min[ µ ( x1 ), µ ( x2 )]
1 x = x
µ x ( x) = 
An unimodal function implies that whenever 0 x ≠ x
µ(x)>µ(y) for a given set A then x is closer to the
ideal definition of A than y.
1
unimodal bimodal
Classical Set Function Linear Functions

Classical set functions are used to define Very simple fuzzy set.
Classical sets.
µ(x) µ(x)
1
x1 x2 x
x1 x
4
Triangular Functions Triangular Functions
Easy to implement, allowing representation of
May be generated from 3 real values.
very complex fuzzy sets.
 0 x<a
( x − a ) /(b − a ) a ≤ x ≤ b
µ(x) 
tri ( x; a, b, c) = 
 (c − x) /(c − b) b ≤ x ≤ c
 0 x>c
a b c x
Trapezoidal Functions Trapezoidal Functions

Easy to implement, allowing representation of
very complex fuzzy sets.
 0 x<a
µ(x) x−a
b − a a≤x<b

trap ( x : a, b, c, d ) =  1 b≤x<c
d − x c≤x<d
d − c
 0 x≥d
x
a b c d
Sigmoid Function Sigmoid Function

It is described by two parameters: the inflexion
point (c), that is the point where the function is
equal to 0.5 and a parameter that defines the
shape.
Sig ( x; a, c) = 
1
−a ( x −c )
1 + e
5
S Function S Function
It is described by three parameters: its 0 degree
value (α), its 1 degree value (γ) and the inflexion
point (β), that is the point where the function is
1
equal to 0.5 (β=(α +γ)/2).
 0 x ≤α
  x −α 
2
 2×   α <x≤β
  γ −α 
S ( x, α , β , γ ) = 
1 − 2 ×  x − α  β < x ≤ γ
2
γ −α 
 0
 1 x>γ α β γ
Beta Function Beta Function

It is defined by two parameters, the value where
Beta−function
the function is equal to one (γ) and the inflexion 1
gamma = 5, beta = 3
(β). 0.9 gamma = 5, beta = 2
gamma = 5, beta = 1
0.8
gamma = 5, beta = 0.5
0.7
0.6
1
B ( x, γ , β ) = 0.5
 x −γ 
2
0.4
1+  
 β 
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10
Support of fuzzy sets Empty Fuzzy Set

The support of a fuzzy set A, defined A fuzzy set (A=Ø) is empty if its
in the universe of discourse X, is the membership function is zero
classical set defined as everywhere in its universe of discourse.
A = ∅ if µ A ( x) = 0, ∀x ∈ X
S A = {x ∈ X | µ A ( x) > 0}
An empty fuzzy set has an empty
support
6
Compact Support Alpha-Cuts Sets
The support is compact when the set is smaller The classical set Aα, called alpha-
alpha-cut set,
set is
than the size of the universe of discourse. the set of elements whose degree of
membership in A is no less than α. It is defined
If the support were not compact then several as:
Aα = {x ∈ X | µ A ( x) ≥ α }

rules would be activated at every input causing
an increase in the system load.
Non Compact Support Compact Support

The classical set A’α is called strong alpha-
alpha-
cut set. It is defined as:
A = {x ∈ X | µ ( x ) > α }
'
α A
Alpha-cut Set Example of Alpha-cuts
Universe of Discourse (Ages):

{5,10,20,…
{5,10,20,…,80}
Young = {(15,0), (17, 0.2), (19, 0.4), (21,0.6), (23,

0.8), (25,1.0), (27, 0.8), (29, 0.6), (31,0.4), (33, 0.2),
(35,0.0)}
Young0.2 = {17, 19, 21, 23, 25, 27, 29, 31, 33}
Young’0.2 = {19, 21, 23,, 25, 27, 29, 31}

31
Alpha-cuts Level Set

The set of all levels alpha that represent
distinct alpha-cuts of a given fuzzy set
A is called a level set of A.
Λ A = {α | µ A ( x) = α for some x ∈ X }
Ex.
Λ young = {0.2, 0.4 ,0.6, 0.8,1.0}
7
Resolution Principle Cardinality
The membership function of a fuzzy set The cardinality |A| of a fuzzy set A is
A can be expressed in terms of its defined as
alpha cuts as (+ is union):
| A |= ∑ µ ( x )
Λ = {α 0 , α1 ,K,α n } x∈ X
A = α 0 × Aα 0 + α1 × Aα1 + L + α n × Aα n The relative cardinality of A is defined

as
α if µ A ( x) ≥ α i
A
A =
µα × Aα =  i X
0 otherwise
0 0
Cardinality - Example Cardinality - cont

Consider the set When the fuzzy set is continuous the
A = {( 6 . 5 , 0 . 25 ), ( 7 , 0 . 5 ), ( 7 . 5 , 0 . 75 ), ( 8 ,1 ), cardinality is defined as
( 8 . 5 , 0 . 75 ), ( 9 , 0 . 5 ), ( 9 . 5 , 0 . 25 )}
defined in the universe of grades from 0 to 10 in A = ∫ µ A ( x)dx

0.5 steps. The cardinality of A is x
| A |= 0.25 + 0.5 + 0.75 + 1 + 0.75 + 0.5 + 0.25 = 4.0
So the relative cardinality of A is

4.0
A = = 0.2
21
Fuzzy Cardinality Fuzzy Cardinality Example

When a fuzzy set A has a finite support, Young = Young = {(15,0), (17, 0.2), (19,
its fuzzy cardinality |Ã| is a fuzzy set 0.4), (21,0.6), (23, 0.8), (25,1.0), (27, 0.8),
(fuzzy number) defined on N whose (29, 0.6), (31,0.4), (33, 0.2), (35,0.0)}
membership function is defined by
µ| A~| (| Aα |) = α ∀α ∈ Λ A
µ| A~| (| Aα |) = α ∀α ∈ Λ A
| y oung |= 0.2 / 9 + 0.4 / 7 + 0.6 / 5 +
~
0 .8 / 3 + 1 .0 / 1
8
Height of a fuzzy set Fuzzy Subsets
The height of a fuzzy set is the highest If the membership grade of each
membership value of its membership element of the fuzzy set A is less than
function or equal its membership grade in fuzzy
set B, then A is called a subset of B.
H A = max x∈X {µ A ( x)}
A ⊆ B if µ A ( x) ≤ µ B ( x ) ∀x ∈ X
A fuzzy set is defined as normal when
HA=1 and subnormal when it is not.
Equal Fuzzy Sets Distance

If the membership grade of each Measures the distance that a value is from the
ideal definition of the set. It is defined as
element of the fuzzy set A is equal its
membership grade in fuzzy set B, then  ∞ µ A ( x) = 0

A is equal to B. d ( A, x) =  1
− 1 µ A ( x) ≠ 0
 µ A ( x)
A = B if µ A ( x) = µ B ( x)
A = B if A ⊆ B and B ⊆ A
Distance Examples of Fuzzy Sets
Distance d(A,x)
1
Set µA(x)
9
Examples of Fuzzy Sets Examples of Fuzzy Sets
Examples of Fuzzy Sets Examples of Fuzzy Sets
10

Universe of Discourse Membership Function: Adriano Joaquim de Oliveira Cruz - NCE e IM, UFRJ Adriano@

Enviado por

Dados do documento

Descrição original:

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Universe of Discourse Membership Function: Adriano Joaquim de Oliveira Cruz - NCE e IM, UFRJ Adriano@

Enviado por

Direitos autorais:

Formatos disponíveis

Summary

Fuzzy Sets - Introduction Classical Sets

@2001 Adriano Cruz NCE e IM - UFRJ Fuzzy Sets Introduction 2

Crips Sets Classical Sets

Classical Sets/Problems Truth Tables

The Laws of Aristotle The Laws of Aristotle

This is called the law of the excluded middle.

Fuzzy Sets Fuzzy Sets - definition

It isn’t probability! It isn’t probability!

Fuzzy Sets Representation 1 Fuzzy Sets Representation 2

Unimodal Functions Singletons

Classical Set Function Linear Functions

Trapezoidal Functions Trapezoidal Functions

Sigmoid Function Sigmoid Function

Beta Function Beta Function

Support of fuzzy sets Empty Fuzzy Set

Non Compact Support Compact Support

Alpha-cut Set Example of Alpha-cuts

Universe of Discourse (Ages):

Young = {(15,0), (17, 0.2), (19, 0.4), (21,0.6), (23,

Young’0.2 = {19, 21, 23,, 25, 27, 29, 31}

Alpha-cuts Level Set

A = α 0 × Aα 0 + α1 × Aα1 + L + α n × Aα n The relative cardinality of A is defined

Cardinality - Example Cardinality - cont

defined in the universe of grades from 0 to 10 in A = ∫ µ A ( x)dx

So the relative cardinality of A is

Fuzzy Cardinality Fuzzy Cardinality Example

Equal Fuzzy Sets Distance

Distance Examples of Fuzzy Sets

Examples of Fuzzy Sets Examples of Fuzzy Sets

Você também pode gostar