B-mine: Frequent Pattern Mining and Its Application to Knowledge Discovery from Social Networks

摘要

As an important data mining task, frequent pattern mining has drawn attention from many researchers. This has led to the development of many frequent pattern mining algorithms, which include Apriori-based, tree-based, and hyperlinked array structure-based algorithms, as well as vertical mining algorithms. Although these algorithms are efficient and popular, they also suffer from some drawbacks. To tackle these drawbacks, we present in this paper an alternative algorithm called B-mine that uses a bitwise approach to mine frequent patterns. Evaluation results show the space- and time-efficiency of B-mine for frequent pattern mining, as well as the practicality of B-mine for social network analysis and knowledge discovery from social networks.