MUTI-KEYWORD SEARCH WITH PRESERVING PRIVACY OVER ENCRYPTED DATA IN THE CLOUD

MUTI-KEYWORD SEARCH WITH PRESERVING PRIVACY OVER ENCRYPTED DATA IN THE CLOUD A.Shanthi 1, M. Purushotham Reddy 2, G.Rama Subba Reddy 3 1 M.tech Scholar (CSE), 2 Asst.professor, Dept. of CSE, Vignana Bharathi Institute of Technology (VBIT),VidyaNagar, Pallvolu, Proddatur, Kadapa(Dist),Andhra Pradesh (India) 3 Working as Associate Professor and Head of Department (CSE), Vignana Bharathi Institute of Technology (VBIT),VidyaNagar, Pallvolu, Proddatur, Kadapa (Dist), Andhra Pradesh (India) ABSTRACT Cloud computing has become fame in the society and more users and more data owners centralize their secure data into the cloud. As there is huge amount of data files that are stores in the cloud server, it is important to provide security for the cloud and for the keyword based search service to data user. Here in order to protect this data, sensitive data is usually encrypted before ending the data to the cloud server, which makes the search technologies on plain text unusable. In this paper, we have to propose a sematic multi-keyword ranked search scheme over the encrypted cloud data, which simultaneously meets a set of privacy requirements. Here we propose the Latent Semantic Analysis to reveal the relationship between the terms and the documents and this latent semantic analysis takes this advantage of the higher-order structure in the terms with document ( semantic structure ) receives a diminished measurement vector space to speak to words and documents. Consequently, the relationship between terms is consequently caught. Besides, our plan utilize secure "k-nearest neighbour (k-nn)" to accomplish secure pursuit usefulness. The proposed plan could return the accurate coordinating documents, as well as the records including the terms inactive semantically related to the inquiry watchword. At long last, the exploratory result exhibits that our technique is superior to the first MRSE plan. I. INTRODUCTION Due to the fast extension of data, the data owner proprietors tends to store their data into the cloud to release the weight of the storing the data and the maintenance. Here the cloud user and the cloud server are not in the same domain; our data may be exposure to the risk. Thus, when the user wants to sends the data to the cloud before that user need to encrypt that sensitive data to protect the data privacy and then after that that encrypted data can be sent directly to the cloud. Fuzzy keywords searches have been developed by chuah et al propose a privacy-aware bed-tree method to support fuzzy multi-keyword search. This multi-keyword uses to edit the distance to build fuzzy keyword sets. Here the bloom filers are used to search the keyword. Here then it builds the index tree for all the data and files where each leaf node a hash value of a keyword. In this paper we have clear the issue of the latent search of the multi keyword latent semantic ranked search over encrypted cloud data and fetching the related files. Here we have proposed a new scheme named the latent semantic analysis(lsa) this is based on the multi keyword rank search which supports the multi-keyword latent 109 P a g e

semantic ranked search. By using LSA technique it will display not only exact matching data or files but also the files including the terms latent semantically associated to the query keyword. Proposed scheme: In this project we have discuss detail information about our schemes. In this process first propose that is Latent Semantic Analysis to implement the latent semantic multi keyword rank searched. II. OUR SCHEMES In our scheme data owner requires outsource n data files, those are {d1,d2,d3, dn} that user prepares to in cloud server in outsource encrypted while still working the capability to search files via cloud server. Now data owner built set of n district elements builds secure searchable index w=extracted from the file collections D. From the above def. Latent Semantic Analysis now data owner built a term document matrix A. Now it can be divided into matrix three other matrices, after we reduce the diversions the true matrix that new matrix A which is calculated the best reduced -dimension that helps approximation to the original term document matrix. That have t keywords of interest in W as given input, one binary vector Q is created in each bit Q[j] indicates whether W=J here W is true or false. The similarity score is expressed as A[j] inner product of data vector Q is query vector A[j] for the j-th column of the matrix A. Set up: now the data user is owner generates.the data owner generates a n+2 bit vector and hide metrics invertible matrices{m1,m2}. The secret key SK is form of a types as {x,m1,m2}. Build Index(A,SK) now the data owner separates a term document matrices A from D and decompose that is split into three matrices according the above scheme now the data owner find r-the statically structure and system latent structure and get clear of the obscuring noise. To minimize the dimensions the K columns of s and then removing related columns U and V respectively after multiple this three matric and get the result work A now consider the privacy it is must should encrypt matrices A before out sourcing after complete the process of extracting of dimensions the original matrices A[j] is extended into (n+2) dimensions instead the result of n. In the results of n numbers the (n+1) th entry is A[j] is set to value random number Ej. And (n+2) th value A[j] set as value 1 When the extending the dimensions, finally A[j] represents. The sub index is is built here. III. TRAP DOOR(W) Now keywords of t interest W as input, one binary vector Q are generated. The (n+1) th entry Q is set to a random number 1.and then scaled by random number r is not equal to zero. The (n+2) th number Q is set to a random number t during the extending of dimensions. Q can be represented as.same process is 110 P a g e

applying encryption for above trap door Tw is generated as to the data user. The final scores will be like this. The top rank id list Dw In the proposed scheme we add some random numbers to the final score which helps clearly display security length. IV. ARCHITECTURE In this architecture we have three modules they are data owner, the client data and the cloud server. Here in this architecture the data owner has a collection of the data documents D={d1,d2,d3,..,dm}. A set of distinct keywords W={w1,w2,w3, wn}is the extracted from the data collection D. The data owner will first build an encrypted searchable index I from the data collection D. Here the data owner will upload the data in the encrypted format and that data will be uploaded in the cloud server. Data users will provide t keywords for the cloud server. The cloud server will take care of the top files or most recent files or relevant data to the search query. The cloud server follows both the designated protocol specification but at the same time analyses data in its storage data and message flows received during the protocol. 4.1 Latent Semantic Search In this we aim to develop the latent semantic relationship between the terms and the documents. Here we use the techniques to estimate the latent semantic structure, and get rid of the obscuring noise. 111 P a g e

4.2 Multi-Keyword Ranked Search Here in this module it supports both multi-keyword query and support result ranking. 4..3 Privacy-Preserving Here in this module we have a designed to meet the security and privacy requirements and protect and prevent the cloud server from learning additional information from index and trapdoor. 4.3.1 Index Confidentiality Here the TF values of keywords are stored in the index. Therefore the index that is stored in the cloud storage or server need to be encrypted. 4.3.2 Trapdoor Unlink ability Here in this module the cloud server should not be able to deduce relationship between the trapdoors. 4.3.3 Keyword Privacy The cloud server will not discern the keyword in query, index by analysing the statistical information like term frequency. V. SINGLE KEYWORD SEARCHABLE ENCRYPTION In the encryption method of traditional single key searchable encryption schemes is works basically build an encrypted searchable index so it is include hidden to the upto granted permissions via trapdoors generated with help of secret keys. It helps rank key word search. Which helps are keyword frequency to rank results it is nothing but undifferentiated results. However it helps only single keyword search in public key settings. In cloud anyone public key can write stored data on server but only recognized users search only private key. public key solutions are normally very expensive, and keyword privacy cloud not be secure in protected in the public key settings, because server could change encrypt and keyword with public key, received help of trap door to find the cipher text. VI. BOOLEAN KEYWORD SEARCHABLE ENCRYPTION To more effective search functionalities, introduce conjunctive keyword search proposed encrypted data in conjunctive keyword. Those schemes are damage large overhead caused by their basic primitives. Such as calculation cost by linear map, cost by secret sharing in communication have more general search methods. In now predicate encryption schemes are introduced recently support both search methods of conjunctive and disjunctive. In search methods of conjunctive it returns the result of all-or-nothing, this means it only submit those documents all key words specified by the query search appear. Another approach is disjunctive keyword search result returns undifferentiated results, It contains every document file that have a subset of special keyword seven only one keyword of interest. In this project proposed inner product queries in predicate encryption only find out whether two vectors orthogonal or not. That is inner product value is related result except is equal to zero. Without find the providing capability to compare related inner product is predicate encryption is not performing ranked search.. 112 P a g e

VII. CONCLUSION In this project An Efficient and Privacy-Preserving Semantic Multi-Keyword Ranked Search over Encrypted Cloud Data is proposed. in middle supports latent semantic search. We use the vector it contains TF values as index to documents. This vector contains a matrix. From this analysis the latent semantic association between documents and terms by latent semantic analysis. Consider the privacy into privacy and security employ a secure splitting NN method encrypted and queried vector. It helps accurate ranked result and protects the confidence data as well as possible. References [1]. Armbrust, M., et al., A view of cloud computing. Communications of the ACM, 2010.53(4):p. 50-58. [2]. Chuah, M. and W. Hu. Privacy-aware bedtree based solution for fuzzy multi-keyword search over encrypted data. in Distributed Computing Systems Workshops (ICDCSW), 2011 31st International Conference on. 2011. IEEE. [3]. Deshpande, S., et al., Fuzzy keyword search over encrypted data in cloud computing. World Journal of Science and Technology, 2013.2(10). [4]. Wang, C., et al. Secure ranked keyword search over encrypted cloud data. in Distributed [5]. Computing Systems (ICDCS), 2010 IEEE 30th International Conference on. 2010. IEEE. [6]. Deerwester, S.C., et al., Indexing by latent semantic analysis. JASIS, 1990.41(6): p. 391-407. [7]. Wong, W.K., et al. Secure knn computation on encrypted databases. in Proceedings of the 2009 ACM SIGMOD International Conference on Management of data. 2009. ACM. [8]. Yang, C., et al. A Fast Privacy-Preserving Multi-keyword Search Scheme on Cloud Data.in Cloud and Service Computing (CSC), 2012 International Conference on. 2012. IEEE. [9]. Powers, D.M. The problem with kappa.in Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics. 2012. Association for Computational Linguistics. AUTHOR DETAILS A.shanthi pursuing M.Tech (CSE) Vignana Bharathi Institute of Technology (VBIT), VidyaNagar,Pallvolu, Proddatur, Kadapa(dist), Andhra Pradesh 516 362 M. Purushotham Reddy received his M.Tech (Computer Science & Engineering) from Jawaharlal Nehru Technology University, Anantapuramu and pursuing Ph.D in JNTUA, Anantapuramu. Presently he is working as Associate Professor in Computer Science & Engineering, Vignana Bharathi Institute of Technology, Proddatur, Kadapa dist, A. P., India. 113 P a g e

G.RamaSubbaReddy received his M.E (Computer Science &Engineering) from Sathyabama University, Chennai.Presently he is working as Associate Professor and Head of the Department in Computer Science & Engineering, Vignana Bharathi Institute of Technology, Proddatur, Kadapa Dist.,A.P, INDIA 114 P a g e