A Novel Approach of Association Rule Hiding Using DBCT (Distortion, Blocking and Cryptographic Technique)

Author Name(s): P. R. S. Naidu, B. Prasanth Kumar, Dr. P. Sateesh, Dr. B. Srinivas, Chandra Sekhar Darapaneni
Author Email: prsn1988@gmail.com


Associative rule hiding is a technique used in hiding sensitive data, during data processing to secure the sensitive association rules generated using association rule mining. Several methods were planned within the literature for hiding sensitive data items. Few apply distributed databases across various sites, few indulged data perturbation, and few utilized clustering and few of them employ data distortion technique. Algorithms supporting this method will follow either of the following two techniques. Hide a particular rule with the help of data alteration technique or hide the principles relying on the sensitivity of the items to be hidden. The proposed perspective dependent on data distortion technique which modifies the position of the sensitive items, yet its support is not at all altered and also used the ideology of representative results to shear the rules initially and then hides those sensitive rules. Experimental results exhibit that proposed method hides lot of rules at a minimum range of database scans in contrast to existing algorithms supporting data distortion technique.


Machine learning is categorized into two types. i) Supervised learning is a data mining task of surmising a function from labeled training data. The training data includes set of training examples. Each example in supervised learning has a set of input objects along with desired output values. ii) While the complication of an unsupervised learning task is for discovering hidden structured data which is neither classified nor labeled. The algorithm has to act on that information without guidance.. A. Data Mining: Data mining is the procedure used for sorting out the huge data sets so that to identify patterns and establish relationships by solving problems with the help of data analysis. Data need to be altered in order that information couldn’t be identified by data mining techniques. Handling such sensitive data is the most important criteria from being accessed by the unauthorized activities. This has led to the disclosure of risks if the data is revealed to outside parties. This perspective has led to the research of hiding the sensitive information within database. B. Privacy preserving data mining: Privacy preserving is regarded as an important concern in association with data mining. This includes protecting independent data and the sensitive information without affecting the efficacy of the data. For preserving the privacy of information one should alter the original database in so that sensitive information shouldn’t get involved in the mining result while the non-sensitive information was obtained. For securing sensitive association rules, privacy preserving data mining includes the concept named “association rule hiding”. Before getting involved in this area, association rule mining technique has to be concerned.


In this paper, we have discussed about how association rules are generated and a new approach to hide sensitive data i.e.(DBCT) a hybrid technique which applies data distortion followed by data Blocking and finally cryptographic technique which is being performed in different configuration machines and generated corresponding performance metrics i,e evaluated space and time complexities. On further work, we can implement some efficient association rule hiding techniques which might give better performance results than DBCT.

109 total views, 1 views today

Download PDF File

About the author: admin