Database Reference
In-Depth Information
sequential pattern mining, constrained pattern mining, and graph mining have been
proposed in the literature.
Frequent pattern mining is a rather broad area of research, and it relates to a wide
variety of topics at least from an application specific-perspective. Broadly speaking,
the research in the area falls in one of four different categories:
•
Technique-centered:
This area relates to the determination of more
efficient
algorithms for frequent pattern mining. A wide variety of algorithms have been
proposed in this context that use different enumeration tree exploration strategies,
and different data representation methods. In addition, numerous variations such
as the determination of compressed patterns of great interest to researchers in data
mining.
•
Scalability issues:
The scalability issues in frequent pattern mining are very
significant. When the data arrives in the form of a stream, multi-pass methods
can no longer be used. When the data is distributed or very large, then parallel or
big-data frameworks must be used. These scenarios necessitate different types of
algorithms.
•
Advanced data types:
Numerous variations of frequent pattern mining have
been proposed for advanced data types. These variations have been utilized in
a wide variety of tasks. In addition, different data domains such as graph data,
tree structured data, and streaming data often require specialized algorithms for
frequent pattern mining. Issues of interestingness of the patterns are also quite
relevant in this context [
6
].
•
Applications:
Frequent pattern mining have numerous applications to other major
data mining problems, Web applications, software bug analysis, and chemical
and biological applications. A significant amount of research has been devoted
to applications because these are particularly important in the context of frequent
pattern mining.
This topic will cover all these different areas comprehensively, so as to provide a
comprehensive overview of this broader area.
This chapter is organized as follows. The next section discusses algorithms for
the frequent pattern mining problem, and its basic variations. Section
3
discusses
scalability issues for frequent pattern mining. Frequent pattern mining methods are
advanced data types are discussed in Sect.
4
. Privacy issues of frequent pattern mining
are addressed in Sect.
5
. The applications are discussed in Sect.
6
. Section
7
gives
the conclusions and summary.
2
Frequent Pattern Mining Algorithms
Most of the algorithms for frequent pattern mining have been designed with the tra-
ditional support-confidence framework, or for specialized frameworks that generate
