An efficient privacy preserving on high-order heterogeneous data using fuzzy K-prototype clustering

摘要

In this paper, we propose a privacy-preserving high order large amount of heterogeneous data using distributed high order fuzzy k-prototype framework named distributed multiple exponential kernel possibilistic fuzzy clustering (MEKPFCM), incorporates kernel fuzzy c-means and possibilistic fuzzy clustering algorithms. The privacy-preserving high order MEKPFCM, cluster the heterogeneous dataset by representing each heterogeneous data object as a tensor. In this paper, the cloud server directly performs clustering over encrypted datasets, while achieving maximum accuracy. The fully homomorphic encryption algorithm (FHE) is utilized to protect the high order large amount of heterogeneous data. Moreover, we design a secure integration of map-reduce into our proposed work, which makes our proposed work enormously appropriate for cloud computing environment. Detailed security analysis and experimental results show that proposed MEKPFCM method can effectively cluster a large amount of heterogeneous data. The experimentation of the proposed technique was carried out using UCI machinery skin dataset and the performance was compared with the previous techniques using accuracy and encryption time. Furthermore, MEKPFCM can cluster bigdata by using the cloud computing technology without disclosing privacy.