Abstract:

Cloud-based data storage and processing framework, which enables users can economically and securely share and handle big datasets with low cost, was proposed. The framework studied data mining algorithms based on the matrix, whose focus was the secure Top-k feature vector algorithm. The algorithm used an iterative processing model and the authorized user interacted with the cloud to achieve the expected result. In this process, the interaction of source matrix and intermediate results were kept confidential. In addition, the security of this algorithm was guaranteed by using Paillier encryption and a random perturbation technique. Experimental results show that the proposed method is scalable to big matrices while requiring low client-side costs.

Key words: cloud computing, data mining, iteration, Paillier encryption, security