Frequent itemset mining is a task that can in turn be used for other purposes such as associative rule mining. One problem is that the data may be sensitive, and its owner may refuse to give it for analysis in plaintext. There exist many privacy-preserving solutions for frequent itemset mining, but in any case enhancing the privacy inevitably spoils the efficiency. Leaking some less sensitive information such as data density might improve the efficiency. In this paper, we devise an approach that works better for sparse matrices and compare it to the related work that uses similar security requirements on similar secure multiparty computation platform.
↧