2010 | OriginalPaper | Buchkapitel
Two-Phase Iterative Algorithm
verfasst von : Aris Gkoulalas-Divanis, Vassilios S. Verykios
Erschienen in: Association Rule Hiding for Data Mining
Verlag: Springer US
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
In this chapter, we present a two–phase iterative algorithm (proposed in [27]) that extends the functionality of the inline algorithm of [23] (Chapter 14) to allow for the identification of exact hiding solutions for a wider spectrum of problem instances. A
problem instance
is defined as the set of (i) the original dataset
$$\mathcal{D_O}$$
, (ii) the minimum frequency threshold
mfreq
that is considered for its mining, and (iii) the set of sensitive itemsets S that have to be protected. Since the inline algorithm allows only supported items in
$$\mathcal{D_O}$$
to become unsupported in
$$\mathcal{D}$$
, there exist problem instances that although they allow for an exact hiding solution, the inline approach is incapable of finding it. The truthfulness of this statement can be observed in the experiments provided in Section 15.4, as well as in the experimental evaluation of [26].