Privacy Preserving Data Mining in Partitioned Databases

concealing Preserving info moing in Partiti mavind DatabasesA survey concealment preserving info mining in horizont everyy softenitioned entropybasesKomal Kapadia, Ms.Raksha Chauhan_______________________________________________________________________________________________________Abstract concealing preserving info mining techniques atomic fleck 18 introduced with the aim of extract the relevant knowledge from the salient amount of selective in contour lineation turn protecting the sensible information at the akin time. The success of entropy mining relies on the availability of gamy quality information. To ensure quality of data mining, effective information overlap amidst organizations becomes a vital requirement in todays society. concealment preserving data mining deals with hiding an individuals elegant identity without sacrificing the usability of data. Whenever we atomic number 18 concerning with data mining, Security is measure issue while extract ing data. Privacy Preserving Data Mining concerns with the security measures of data and provide the data on demand as well as amount of data that is required.Index Terms data mining, cover preserving, ECC cryptography, randomized chemical reaction technique._______________________________________________________________________________________________INTRODUCTIONData mining techniques have been widely determinationd in some(prenominal) areas especi comp permitelyy for strategic decision making. The main threat of data mining is to security and concealing of data residing in large data stores. Some of the information considered as private and secret rump be bought out with shape upd data mining tools. Different research efforts are under way to address this trouble of privateness preserving and preserving security. The privacy border has wide range of different meanings. For example, in the context of the health damages accountability and portability act privacy eclipse , privacy means the individuals ability to control who has the access to personal health care information. In organization, privacy means that it involves the definition of policies stating which information is collected, how it is used, how customers are involved and aware in this process. We can considering privacy as Individuals go for and ability to keep certain information about themselves hidden from differents. Privacy preserving data mining refers to the area of data mining that seeks to safeguard fond information from unsolicited disclosure. Historically, issues related to PPDM were first examine by the interior(a) statistical agencies interested in collecting private social and frugal data, such as census and tax records, and making it available for epitome by creation servants, companies, and researchers. Building accurate socio-economical innovativeels is vital for business training and public policy. Yet, there is no way of knowing in advance what advancedel s may be needed, nor is it feasible for the statistical agency to perform all data processing for everyone, playing the role of a indisputable deuce-ace company. Instead, the agency provides the data in a sanitized form that allows statistical processing and protects the privacy of individual records, solving a problem known as privacy preserving data publishing. There are many methods for preserving the privacy. In this surevey many methods try to compute the answer to the mining without reveal any additional information about user privacy.Progress in scientific research depends on the sharing and availability of information and ideas. scarcely the researchers are mainly foc utilise on preserving the security or privacy of individuals. This issue leads to an emerging research area, privacy preserving data mining. For privacy preserving data mining, many authors proposed many technologies. The main aim of this paper is, to develop expeditious methodology to find privacy preserv ing.LITERATURE SURVEYWe have studied some of the related work for the privacy preserving in horizontally partitioned databases. active work for privacy preserving in horizonatally partitioned database has different types of techniques.TYPES OF PRIVACY PRESERVING TECHNIQUESSemi secure companionshipWithout trusted partyWith trusted partyIn without trusted party individually party willinging calculate their own incomplete livelihood and add their own random number and delegates the result to the future(a) party in the ring so that the early(a) party will never know the result of former(a)s and in last the initiator party will disclose the result that is global support.In trusted party each party will calculate their partial support and circularise to the trusted party and add the own random number and send to the next coming localise in the ring so that other party will never know the result of other parties aft(prenominal) that trusted party will disclose the result and send to all internet sites that presents in the ring.Fig. 1 Framework of privacy preserving data mining5SECURE MULTIPARTY dialogueApproximately all Privacy Preserving data mining techniques rely on Secure multi party communication protocol. Secure multi party communication is outlined as a counting protocol at the last part of which no party involved knows anything else except its own inputs the outcome, i.e. the view of each party during the execution can be effectively simulated by the input and output of the party. Secure multi party communication has commonly intemperate on two forward-lookingels of security. The semi-honest model as jointes that every party follows the rule of the protocol, nevertheless is free to later use what it sees during execution of the protocol. The malicious model assumes that parties can arbitrarily cheat and such cheating will non agree moreover security or the outcome, i.e. the results from the malicious party will be correct or the malici ous party will be detected. closely of the Privacy Preserving data mining techniques assume an intermediate model, Preserving Privacy with non-colluding parties. A malicious party May dishonest the results, but will not be able to learn the private data of other parties without colluding with another(prenominal) party.(1)MHS algorithmic program FOR HORIZONTALLY PARTITION DATABASEM. Hussein et al.s Scheme (MHS) was introduced to make better privacy or security and try to reduce communication monetary value on increasing number of sites. Behind this main idea was to use effective cryptosystem and rearrange the communication path. For this, two sites were discovered. This algorithm works with minimum 3 sites. One site acts as Data Mining provoker and other site as a Data Mining Combiner. Rests of other sites were called client sites. This scenario was able to decrease communication time. Fig. shows MHS algorithm.The working of the algorithm is as followsThe initiator generates RS A public tell apart and a private key. It sends the public key to combiner and all other client sites.2. Each site, except initiator computes habitual accompanimentset and local support for each buy at itemset utilise Local Data Mining .3. All Client sites encrypt their computed data using public key and send it to the combiner.4. The combiner merges the received data with its own encrypted data, encrypts it again and sends it to initiator to find global standoff rules.5. Initiator decrypts the received data using the private key. Then it merges its own local data mining data and computes to find global results.6. Finally, it finds global association rules and sends it to all other sites.Fig.2 MHS algorithm11(2) EMHS ALGORITHM FOR HORIZONTALLY PARTITION DATABASEEnhanced M. Hussein et al.s Scheme (EMHS) was introduced to alter privacy and reduce communication cost on increasing number of sites. This algorithm also works with minimum 3 sites. One site acts as Data Mining Initia tor and other site as a Data Mining Combiner. Rests of other sites were called client sites . But this algorithm works on the concept of MFI (Maximal ordinary Itemset) rather of Frequent Itemset.a) MFI (Maximal Frequent Itemset) A Frequent Itemset which is not a subset of any other frequent itemset is called MFI. By using MFI, communication cost is reduced .b) RSA (Rivest, Shamir, Adleman) Algorithm one of the widely used public key cryptosystem. It is based on retentiveness factoring product of two large prime numbers secret. happy chance RSA encryption is tough.(3)MODIFIED EMHS ALGORITHM FOR HORIZONTALLY PARTIOTION DATABASEIn this technique, they used modified EMHS algorithm for improving its efficiency by using Elliptic curve cryptography. here(predicate) Elgamal cryptography technique is used which is of ECC for homomorphic encryption.ELLIPTIC CURVE codingElliptic curve cryptography provides public cryptosystem based on the descrete log problem over integer modulo a prime. Elliptic curve cryptosystem requires a great deal shorter key length to provide a security level same as RSA with larger key length. In this elgamal cryptography is used.ELGAMAL CRYPTOGRAPHYa)A wishes to mass meeting message M with B9.b) B first chooses Prime Number p, reservoir g and private key x.c)B computes its Public Key Y = gx mod p and sends it to A.d) Now A chooses a random number k.e) A calculates one time key K = Yk mod p.f) A calculates C1 = gk mod p and C2 = M*K mod p and sends (C1,C2) to B.g) B calculates K = C1x mod ph) B calculates K-1 = inverse of K mod p i)B recovers M = K-1 * C2 mod pj) Thus, Message M is exchanged between A and B securely.In this system, Elgamal cryptography paillier cryptosystem is used. Here, Elgamal cryptography is used for security purpose. Compared to EMHS algorithm here performance is better in terms of computation time.RANDOMIZED RESPONSE TECHNIQUEIn this technique, here mainly focus on CK secure sum in randomized response technique f or privacy preserving. Here, the multi party transaction data who discover frequent item sets with minimum support.In the randomized response technique, consider the data sets I = I1, I2, I3In and the random number or noise part are denoted by R= R1, R2, R3 Rn, the new set of records are denoted by I1+R1, I2+R2 .In+Rn and after that take a partial support Pij = Pi1, Pi2..Pin so that partial support isP ij=I+R10I=Pij-R10In randomized response secure sum technique, secure sum each site will determine their own data value and send to forerunner site that near to pilot program site and this goes on till the original site collects all the value of data after that the parent site will determine the global support.CK SECURE SUM ALGORITHM10Step1-Consider parties P1, P2, P3Pn.Step2-Each party will generate their own random number R1, R2.RNStep3-Connect the parties in the ring (P1, P2, P3PN) and let P1 is aprotocol initiator.Step4-Let RC=N, and Pij=0 (RC is round counter and Pij is partial support)Step5-Partial support P1 site calculating by using following formulaPsij = Xij.support Min support * DB + RN1 RNnStep6-Site P2 computes the PSj for each item received the list using theList using the formula,PSij= PSij + Xij. Support minimum support * DB +Rn1-Rn (i-1)Step7-While RC =0 begin for j=1 to N dobegin for I=1 to N doStep8-P1 exchange its position to P (j+1) mod N andRC=RC-1endStep9-Party P1 allowance the result PijStep10-EndIn ck secure sum technique, mainly focused on for computing global support inside a scenario of homogeneous database and provides the high security to the database and hacking of data is zero.CONCLUSIONIn this paper we reviewed five privacy preserving technique in horizontally partitioned database. In MHS algorithm RSA cryptography is used. In EMHS algorithm, by using MFI approach true statement is high compared to MHS. Modified EMHS algorithm used elgamal technique so privacy is high than EMHS technique. Randomized response technique provid es high security to the database compared to other techniques. In future we can compute less number of rounds instead of n number of rounds. Here , we can use encryption technique for encrypting random number and sends it to the predecessor.REFERENCES1 Neelamadhab Padhy, Dr. Pragnyaban Mishra Rasmita Panigrahi. The Survey of Data Mining Applications and Feature Scope. 2012 IJCSEIT.2 Xinjun qi, Mingkui zong. An overview of privacy preserving data mining. 2011 ICESE.3 Kishori pawar, Y.B. gurav. Overview of privacy in horizontally distributed databases. 2014 IJIRAE.4 Manish Sharma, Atul chaudhary , Manish mathuria Shalini chaudhary. A review canvass on the privacy preserving data mining techniques and approaches.. 2013 IJCST.5 Shweta taneja, shashank khanna, sugandha tilwalia, ankita. A review on privacy preserving data mining techniques and research challenges. 2014 IJCSIT.6 Jayanti dansana, Raghvendra kumar Jyotirmayee rautaray. Techniques for privacy preserving association rule mining in distributed database. 2012 IJCSITS.7 Xuan canh nguyen, Tung anh cao. An enhanced scheme for privacy preserving association rules minig on horizonatally distributed databases. 2012 IEEE.8 Manish Sharma, Atul chaudhary, Manish mathuria, Shalini chaudhary Santosh kumar. An efficient approach for privacy preserving in data mining. 2014 IEEE.9 Rachit v. Adhvaryu, Nikunj h. Domadiya. Privacy preserving in association rule mining on horizontally partitioned database. 2014 IJARCET.10 Jayanti Dansana , Raghvendra Kumar , Debadutta Dey. Privacy preservation in horizontally partitioned databases using randomized response technique. 2013 IEEE.11 Rachit v. Adhvaryu, Nikunj h. Domadiya, investigate Trends in Privacy Preserving in tie-up Rule Mining (PPARM) On Horizontally Partitioned Database. 2014 IJEDR.12 Agrawal D. Aggarwal C. C. On the Design and Quantification of Privacy-Preserving Data Mining Algorithms.ACM PODS Conference, 2002.13 D.W.Cheung,etal.,Ecient Mining of Association Rules in Distributed Databases, IEEE Trans. Knowledge and Data Eng., vol. 8, no. 6, 1996,pp.911-922.