Home

itemsetsthose

Itemsetsthose is a theoretical construct used in the field of pattern mining to describe a class of selection operators applied to collections of itemsets. The term combines itemsets with a demonstrative element to emphasize the act of choosing those itemsets that meet a specified condition. While not a universally standardized term, it appears in some theoretical discussions as a convenient way to describe intersection- or predicate-based filtering of itemsets.

Definition and notation: Let S be a set of itemsets drawn from a universe of items. Let

Example: Consider S = { {bread, milk}, {bread}, {milk, cheese}, {butter} } and T = {milk, butter}. Then itemsetsthose(S, T)

Applications: itemsetsthose can be used to filter results in data mining workflows, such as extracting itemsets

Relation to related concepts: The operator is akin to a selection or filter operation in relational algebra

See also: frequent itemset, closed itemset, association rule, pattern mining.

---

P
be
a
predicate
that
can
be
evaluated
on
an
itemset
s
∈
S.
The
operator
itemsetsthose(S,
P)
yields
the
subset
{
s
∈
S
|
P(s)
is
true
}.
In
a
common
specialization,
where
P
asks
for
a
nonempty
intersection
with
a
reference
set
T,
the
operator
becomes
itemsetsthose(S,
T)
=
{
s
∈
S
|
s
∩
T
≠
∅
}.
=
{
{bread,
milk},
{milk,
cheese},
{butter}
}.
that
involve
a
particular
category
(for
example
dairy
products)
or
preparing
input
for
downstream
tasks
like
association
rule
mining,
clustering,
or
visualization.
and
is
related
to
intersection-based
filtering
in
frequent
itemset
mining.
It
is
independent
of
how
the
initial
itemsets
were
produced
and
can
be
applied
to
any
suitable
collection.