Introduction - Association Rule Hiding for Data Mining - page 7

1.4 Summary
Since its first appearance as a problem in 1999, association rule hiding has been
extensively studied by the data mining research community, leading to a large body
of significant research work over the past years. The approaches that have been
developed span from simple, time-efficient and memory-efficient heuristics that se-
lect transactions and items to sanitize, to more complicated and sophisticated so-
lutions which conceive the hiding process as an optimization problem and solve
it by using specific optimization techniques. The following chapter provides the
background along with the necessary terminology for the formal definition of the
problem. Next, in Chapter 3, we present a brief taxonomy of the various association
rule hiding methodologies that have been proposed over the years, along four or-
thogonal dimensions. Each class of approaches is further presented in detail in the
corresponding part of the topic. Specifically, heuristic methodologies are covered in
the second part, while the third part deals with border-based approaches. The most
recent class of approaches that involves exact hiding methodologies is presented in
the fourth part of the topic. Finally, Chapter 4 discusses some methodologies for
sensitive knowledge hiding in research areas related to that of association rule hid-
ing and in particular in the areas of classification, clustering and sequence hiding.
Last, Chapter 5 summarizes the first part of the topic.