K-Nearest Neighbor Categorization on Secure Data Access in Cloud

For the last few years, a extensive research has been going on query processing of relation data and more practical and theoretical solution have been suggested to query processing under different scenarios. Now days cloud computing technology is increasing rapidly, so users now have the chance to store their data in remote location. However, different privacy issues are raised on cloud computing, important data needs to be encrypted before store the data on cloud storage. In extra, query processing methods have to be supported by cloud storage; otherwise, there is a no chance to store data on remote location of cloud storage. To perform the operation by queries on encrypted data without the decrypting by cloud is an important challenging issue. In our proposed system we take focus for resolving the k-nearest neighbor (kNN) query issues over the encrypted outsourced data on cloud storage: user issues of encrypted query information to cloud storage and return the k closest information to user by cloud. We propose k-nearest neighbor protocol that protects the input query of user, confidentiality of data and access pattern of data. Also we examine our protocol efficiency by different experiments. However, as stated above Privacy Preserve k-nearest neighbor (PPkNN) is composite issues and it cannot be achieved straightly by method of the existing k-nearest neighbor techniques on encypted data. We improve our proposed system and produce new solution for Privacy Preserve k-nearest neighbor (PPkNN) classifier issues on encrypted data.