K-NN CLASSIFICATION OVER SECURE ENCRYPTED RELATIONAL DATA IN OUTSOURCED ENVIRONMENT

Abstract

Data Mining has wide applications in numerous zones for example banking, medicine, and exploratory examination and among government offices. Classification is one of the ordinarily utilized assignments as a part of data mining applications. For as long as decade, because of the ascent of different security issues, numerous hypothetical and handy answers for the arrangement issue have been proposed under diverse security models. On the other hand, with the later prevalence of distributed computing, clients now have the chance to outsource their information, in encoded structure, and additionally the information mining assignments to the cloud. Since the information on the cloud is in encoded structure, existing security safeguarding characterization procedures are not persistent. In this paper, we concentrate on taking care of the arrangement issue over encoded information. Specifically, we propose a protected K-NN classifier over encoded information in the cloud. The proposed convention ensures the secrecy of information, security of client's information question, and conceals the information access designs. To the best of our insight, our work is the first to add to a safe K-NN classifier over encoded information under the semi-fair model. Likewise, we observationally investigate the proficiency of our proposed convention utilizing a real world dataset under diverse parameter settings.