An Implementation of Privacy Preserving “IF THEN ELSE” Rules for Vertically Partitioned Data

Abstract

The multiparty data utilized with data mining techniques securely to prevent privacy of end-user is known as the privacy-preserving data mining. In this paper, a technique for mining “IF THEN ELSE” rules for the decision-making process is proposed for the privacy-preserving data mining environment. In this context, it is assumed that each participant has a different set of attributes and a common class label. Additionally, not a single party wants to disclose the data contents. Additionally, each party wants to recover its own part or contributed part of information recovery. Therefore, AES and SHA1-based cryptographic algorithms are used for preventing the sensitive amount of data. Additionally to ensure the privacy, the data is ciphered at the client end. In addition to that the C4.5 decision tree algorithm is used for processing the data and extraction of “IF THEN ELSE” rules. The implementation of the proposed technique is provided herein JAVA technology. Additionally, the evaluation of the proposed technique is given here in terms of accuracy, error rate, and memory and time usages. Finally to justify the efforts, the normal dataset (without encryption) is used with a C4.5 decision tree to measure the utility of published decisional rules.