Distributed Mining Reducts of Attributes on Horizontally Partitioned Data

The field of distributed data mining (DDM) has emerged as an active area in recent years because the key challenge in knowledge discovery is the extraction of knowledge from massive databases. Rough set theory (RST) is one of the powerful approaches in data mining, which has been demonstrated to have its usefulness in successfully solving a variety of problems. But there is almost no literature related to the distributed computation in RST. In this paper, the relation between the reducts of partitioned data and global data are discussed. An useful proposition is obtained, which shows that every reduct of global data determinedly has subsets as the elements in reducts of partitioned data. In following, two algorithms, DMR and PPDMR, are proposed for distributed mining of reducts on horizontally partitioned data. DMR concerns the reduction of time complexity while PPDMR focuses on privacy preserving. Experiments and propositions show the excellent function of DMR and PPDMR through practical and academic ways. Just because the pivotal status of reduct in RST, the algorithms proposed in this paper will show good foreground in distributed data mining.