Scanning once a large distributed database to mine global association rules by growing a prefix tree for each local transaction

Scanning once a large distributed database to mine global association rules by growing a prefix tree for each local transaction F. Wang & N. Helian Department of Computing, London Metropolitan University, UK Abstract Most of the popular data mining algorithms are designed to work for centralized data and they often do not pay attention to the resource constraints of distributed and mobile environments. In support of the third generation of data mining systems on distributed and massive data, we proposed an efficient distributed and mobile algorithm for global association rule mining, which does not need to ship all of local data to one site thereby not causing excessive network communication cost. In this algorithm the contribution from each transaction is comprehensively taken into account by growing a prefix tree for each transaction and enumerating all subsets of the transaction itemset. There is no need at all to store and re-scan the previously-scanned transactions, which will be discarded......