A New Method for Controlling the Side Effects in Hiding Association Rules
Farsad Zamani Boroujeni, Maryam Nafar Dehsorkhi · Majlesi journal of energy management/Majlesi journal of mechatronic systems · 2016
Today, privacy preserving has been turned into a new research branch in the data mining in transactional databases. Despite the great advances that have been achieved in this area, most guidelines are powerless in the management of side effects of changes in data. This paper offers an algorithm for optimal control of side effects in the field of privacy preserving in the category of discovery of association rules. The proposed algorithm tries to hide sensitive association rules with the control of selected side effects by an interactive mechanism for selecting side effects. The emphasis of algorithm on the removal or insertion of an element per each selective transaction has the greatest influence on the sensitive rules and the least effect on the other rules (non-sensitive). Experimental results show that the proposed algorithm cause to control effectively the side effects and also reduce the number of lost rules compared with existing solutions.