Privacy-Preserving Outsourced Collaborative Frequent Itemset Mining in the Cloud
Big Data management and analytics has revolutionized the way how organizations collect, store, process and retrieve, huge volumes of data. In order to fully leverage the potential of big data, it is often that organizations need to collaborate and analyze their combined data, and thus, improving the accuracy of results. However, due to government regulations and internal privacy policies, organizations cannot freely share their data with one another. Existing secure multiparty computation techniques along this direction are very expensive. In this paper, we develop a protocol that facilitates multiple users to outsource their encrypted databases as well as the frequent itemset mining task to a cloud environment in a collaborative and privacy-preserving manner. Our solution is built using the well-known apriori algorithm in order to boost the performance of frequent itemset mining in the cloud. Our comprehensive analysis has demonstrated that the proposed solution preserves the confidentiality of participating users. Additionally, our solution ensures that the entire frequent itemset mining task is performed on the cloud-side, thereby fully utilizing the cloud computing services to handle big data needs and incurring negligible cost on the end-users.
MSU Digital Commons Citation
Samanthula, Bharath Kumar, "Privacy-Preserving Outsourced Collaborative Frequent Itemset Mining in the Cloud" (2018). Department of Computer Science Faculty Scholarship and Creative Works. 492.