A Hybrid Algorithm based on Heuristic Method to Preserve Privacy in Association Rule Mining

Abstract

By developing technology and competition in different fields, preserving sensitive data is considered as a problematic issue for users. As long as users do not need to share their data, they preserve them in different ways, such as encryption and hiding them in personal devices like cell phones and computers. When users find it necessary to share their personal data, privacy preserving data mining will help them. In the present study, we introduce two algorithms called ISSDD (Intelligent Selection of Sanitization in Dense Database) and ISSSD (Intelligent Selection of Sanitization in Sparse Database) to decrease side effects such as hiding failure, losing non-sensitive rules, making new rules and also hiding sensitive rules without any restriction in the number of items in the left and right hand. In the suggested algorithms distortion technique is used to hide sensitive rules by declining confidence-based and support-based of rules.