I.J. Information Engineering and Electronic Business, 2018, 3, 45-50 Published Online May 2018 in MECS (http://www.mecs-press.org/) DOI: 10.5815/ijieeb.2018.03.06 Copyright © 2018 MECS I.J. Information Engineering and Electronic Business, 2018, 3, 45-50 A Novel Algorithm for Association Rule Hiding T.Satyanarayana Murthy 1 , N.P.Gopalan 2 Department of Computer Applications, National Institute of Technology, Tiruchirapalli Email: 1 murthyteki@gmail.com, 2 npgopalan@gmail.com Received: 28 August 2017; Accepted: 08 January 2018; Published: 08 May 2018 Abstract—Current days privacy concern about an individual, an organization and social media etc. plays a vital role. Online business deals with millions of transactions daily, these transactions may leads to privacy issues. Association rule hiding is a solution to these privacy issue, which focuses on hiding the sensitive information produces from online departmental stores ,face book datasets etc..These techniques are used to identify the sensitive rules and provide the privacy to the sensitive rules, so that results the lost rules and ghost rules. Algorithms developed so far are lack in achieving the better outcomes. This paper propose two novel algorithm that uses the properties from genetic algorithm and water marking algorithm for better way of hiding the sensitive association rules. Index Terms—Privacy, privacy preserving, lost rules, ghost rules I. INTRODUCTION Online business to deal with millions of transactions. Privacy for the sensitive information is a major challenging task. Association rule mining [1,2,3] a major technique for market basket analysis where processing the transactions and generate the association among these transactions. These association are based on the parameters like support and confidence. support determines the occurence of the item appears in the transactional dataset, where as confidence determines the strength of the rule. These rules are divided into SR and NSR based on the minimum support and maximum confidence parameters. This article focuses on hiding the sensitive rules. During this hiding process ,hiding failure is a major parameter determines the failure of hiding the sensitive association rules. In this paper, identify the privacy breach while hiding the sensitive association rules. Our goal is reduce the ghost rules and increasing the identification of lost rules. The objective of the article takes the dataset as an input and applies the Aprior rule miner for generating the associations rules among the dataset. Instead of Aprior Algorithm using the Association rule hiding algorithm generates a Sanitized Dataset. DSRRC approach cannot handle ,hiding association rules with multiple items in. The efficiency of MDSRRC algorithm can be poor, so it must require future enhancements. Genetic Algorithm [4,5] is an evolutionary algorithm evolves from the lives of the human beings. Genetic algorithm is made up of collection of individuals called chromosomes. These chromosomes are used to represents the population to develop a solution for the suitable problem. Each chromosome represented with a binary values either 0 or 1. The basic theory behind the genetic algorithm was the survival of fittest proposed by Darwin. The species that live longer can have more fitness leads to more survival less fitness leads to less survivals. The Genetic Algorithm Approach begins with random generation of individuals. These individuals collection together called as an population. During the genetic Algorithm process a new population replaces the old population in each iteration. The best chromosome in the population are chosen. The population will transform into future chromosomes basing on the fitness function. The operations that are performed are initialize the population, selection of chromosomes, calculate the fitness values based on the random values. Association rule mining used for analyze the transactional database for identifying strong rules. This was proposed by Agrawal and Cheung for transactional data-sets. The main purpose is finding the frequently used item sets and generate association rules on the dataset. Let a Transactional Database D consists of t1,t2,t3,t4,t5...tn where T is a collection of items like i1,i2,i3,...in. Support of X->Y determines the the ratio of the records which contains XUY with the D, where D equals number of transactions. Confidence gives the strength of the rules.. II. LITERATURE SURVEY Heuristic, Border, Exact, Reconstruction based and Cryptographic based techniques are used for hiding the sensitive association rules. V. Verykios team [6] mainly focused on hiding the sensitive rules using a cyclic algorithmic approach so that limiting disclosure of sensitive rules. Elena Dasseni, Vassilios S. Verkios,Ahmed K. Elmagarmid,Elisa Bernito [7] proposed methods based on confidence and support. V. S. Verykios, A. K. Elmagarmid, E. Bertino, Y. Saygin, and E. Dasseni [8] proposed five algorithms to hide the sensitive knowledge. Stanley R. M. Oliveira, Osmar R. Zaiane [9] improved the balancing factor for achieving the better privacy results. ADSRCC and RRLR algorithm are used for hiding the sensitive rules was proposed by Komal Shah, Amit Thakkar,Amit Ganatra [10].ADSRRC algorithm was popular algorithm to hide the sensitive