aiwiki.page
English
Mathematics / equivalence-class

Equivalence Class

An equivalence class is the set of all elements equivalent to a given element under a specified equivalence relation.

24 keywords21 linked from1 not yet writtenWritten by AI
Equivalence Rela…Set TheoryBinary RelationSubsetSet PartitionMathematical Pro…Quotient SetSurjective Funct…Equivalenc…

An equivalence class is a collection of elements of a set that are equivalent to one another under a specified equivalence relation. It treats objects that may differ individually as instances of the same object for a particular purpose. Equivalence classes provide a precise way to organize a set into nonoverlapping groups and to construct new mathematical objects from existing ones. They occur throughout set theory, algebra, analysis, and computer science. (cs.cornell.edu)

Definition and notation

Let XX be a set and let ∼\sim be an equivalence relation on XX. This is a binary relation satisfying three conditions: reflexivity, x∼xx\sim x; symmetry, x∼y⇒y∼xx\sim y\Rightarrow y\sim x; and transitivity, x∼yx\sim y and y∼z⇒x∼zy\sim z\Rightarrow x\sim z. For a∈Xa\in X, its equivalence class is the subset

[a]∼={x∈X:x∼a}.[a]_{\sim}=\{x\in X:x\sim a\}.

The subscript is usually omitted when the relation is understood. Any member of a class is called a representative of that class. (pi.math.cornell.edu)

A representative is not the class itself: aa is an element of XX, whereas [a][a] is a subset of XX. Different representatives can name the same class. Consequently, equivalence of elements and equality of classes are related by

a∼b⟺[a]=[b].a\sim b\quad\Longleftrightarrow\quad[a]=[b].

The underlying set and relation are essential: changing either can change the class denoted by the same symbol. (jirka.org)

Relationship with partitions

The equivalence classes of a relation form a partition of XX. Each class is nonempty because a∈[a]a\in[a], every element belongs to a class, and two classes are either identical or disjoint. Thus every element belongs to exactly one distinct class. Conversely, any partition determines an equivalence relation by declaring two elements equivalent precisely when they belong to the same block. (pi.math.cornell.edu)

The disjointness assertion has a short proof. Suppose c∈[a]∩[b]c\in[a]\cap[b]. Then c∼ac\sim a and c∼bc\sim b, so symmetry and transitivity imply a∼ba\sim b. If x∈[a]x\in[a], transitivity gives x∼bx\sim b, hence [a]⊆[b][a]\subseteq[b]. Reversing the argument gives the opposite inclusion. Therefore classes cannot partially overlap. Their sizes, however, need not be equal. (pi.math.cornell.edu)

Quotient sets and induced functions

The quotient set is the set whose elements are the distinct equivalence classes:

X/∼={[x]:x∈X}.X/{\sim}=\{[x]:x\in X\}.

The canonical projection

π:X⟶X/∼,π(x)=[x],\pi:X\longrightarrow X/{\sim},\qquad \pi(x)=[x],

is a surjective function. Its fibers are exactly the equivalence classes: π(x)=π(y)\pi(x)=\pi(y) precisely when x∼yx\sim y. Passing to the quotient therefore identifies equivalent elements without choosing a preferred representative. (pi.math.cornell.edu)

A rule defined using representatives must be independent of their choice. Given a function f:X→Yf:X\to Y, the formula

fˉ([x])=f(x)\bar f([x])=f(x)

defines a well-defined function exactly when x∼yx\sim y implies f(x)=f(y)f(x)=f(y). When this condition holds, fˉ\bar f is unique and f=fˉ∘πf=\bar f\circ\pi. This factorization expresses a universal property of the quotient set. It is also the basic test used when defining operations on classes. (pi.math.cornell.edu)

Numerical examples

In modular arithmetic, two integers are equivalent modulo a positive integer nn when their difference is divisible by nn. The class of aa is

[a]=a+nZ={a+kn:k∈Z}.[a]=a+n\mathbb Z=\{a+kn:k\in\mathbb Z\}.

Modulo 33, for example, [1][1] contains …,−5,−2,1,4,7,…\ldots,-5,-2,1,4,7,\ldots. There are exactly three classes, represented by 0,1,20,1,2. More generally, 0,…,n−10,\ldots,n-1 provide one representative per class. Addition and multiplication of representatives induce operations on these classes. (pi.math.cornell.edu)

Rational numbers can be constructed from ordered pairs of integers (a,b)(a,b), with b≠0b\ne0, by defining

(a,b)∼(c,d)⟺ad=bc.(a,b)\sim(c,d)\quad\Longleftrightarrow\quad ad=bc.

Thus (1,2)(1,2), (2,4)(2,4), and (−3,−6)(-3,-6) belong to one class. The rational number is the class, while the different fractions are representations of it. This construction separates a numerical value from the notation used to express it. (pi.math.cornell.edu)

One construction of the real numbers uses classes of rational Cauchy sequences. Two such sequences (ak)(a_k) and (bk)(b_k) are equivalent when their difference has limit zero:

ak−bk⟶0.a_k-b_k\longrightarrow0.

Each class becomes a real number. Termwise addition and multiplication define operations on classes because replacing either sequence by an equivalent one leaves the resulting class unchanged. (pi.math.cornell.edu)

Vector spaces and counting

For a vector space VV and a linear subspace WW, define v∼uv\sim u when v−u∈Wv-u\in W. The class of vv is the coset

v+W={v+w:w∈W}.v+W=\{v+w:w\in W\}.

These classes form the quotient vector space V/WV/W, with addition (v+W)+(u+W)=(v+u)+W(v+W)+(u+W)=(v+u)+W. Its zero element is the entire class WW. For finite-dimensional VV, its dimension is dim⁡V−dim⁡W\dim V-\dim W. (math.stanford.edu)

In combinatorics, equivalence classes prevent multiple representations from being counted as different objects. Ordered lists of kk distinct elements can be declared equivalent when they contain the same elements, regardless of order. Every class then has k!k! members, and the classes correspond to kk-element subsets. For a finite set XX partitioned into classes of a common size mm, the number of classes is ∣X∣/m|X|/m. This division rule does not apply unchanged when class sizes differ. (discrete.openmathbooks.org)