Privacy Preserving Association Rules by Using Greedy Approach

[attachment=16378]

Introduction
In the data mining, the most well-known method is the
Apriori algorithm. Since the Apriori algorithm is costly to find
candidate itemsets, there are many variants of Apriori that
differ in how they check candidate itemsets against the
database [4]. For example, Zaki et al. [9] presented the
algorithms MaxEclat and CHARM for identifying maximal
frequent itemsets. Furthermore, in [8] employed an efficient
data structure FCET and a novel algorithm GMAR to mine
generalized association rules.
Recently

Problem Formulation
Let Γ={i1,i2,i3…in} be a set of literals, called items. Let
D={T1,T2,…Tm} be a set of transactions which is the database
that is going to be disclosed. Each transaction Tj∈D is a subset
of Γ of itemsets, such that T⊆ Γ. For each transaction Tj is
associated with a unique identifier, which we call TID. We
assume that the items in a transaction or an itemset are sorted
in lexicographic order.


The Framework for Privacy Preservation
As depicted in Figure 2, our framework consists of an
original database, mining algorithm, FCET index tree,
sanitized algorithm used for hiding restrictive association rules
from the database, and a transaction retrieval procedure which
is included in sanitizing algorithm for fast retrieval of
transactions.


The Greedy Approach
As depicted in Figure 3, our greedy approach include the
sanitize procedure, we also use extra two sets that are artificial
set and missing set where the values in sets store for artificial
rules and missing rules, respectively. Initially, the sensitive
rules are stored in the sensitive set. The missing set and
artificial set are empty