K-nearest neighbor classification over semantically secure encrypted relational data

Abstract

Data Mining has spacious applications in numerous areas such as banking, medicine, scientific investigate and among government agencies. Classification is a one of the normally used responsibilities in data mining applications. For the past decade, due to the increase of various isolation issues, many academic and useful solutions to the arrangement problem have been planned under dissimilar security models. However, with the recent popularity of cloud computing, users now have the chance to subcontract their data, in encrypted form, as well as the information withdrawal errands to the cloud. Since the data on the shade is in encrypted form, obtainable privacy-preserving classification techniques are not appropriate. In this paper, we center on solving the categorization difficulty over encrypted data. In scrupulous, it's suggest a secure k-NN classifier over encrypted statistics in the obscure. The proposed protocol protects the confidentiality of data, seclusion of user's input doubt, and hides the data call patterns. To the best of our knowledge, our work is the first to develop a protected k-NN classifier over encrypted data below the semi-honest model. Also, it's empirically examining the competence of our prospect procedure by a real-world dataset below different constraint settings.