A dichotomous algorithm for association rule mining

The analysis of large amounts of data requires important computing resources that may not be available, even in current environment, and there are traditionally two main ways for solving this problem. The first is to use multiprocessor machines, and the second is to use computer clusters. The main drawbacks of these solutions are the expensive cost of machines and their specific utilization. In order to avoid these drawbacks, distributed algorithms for mining association rules have been proposed. However, these algorithms either run high-synchronous methods or lack flexibility to be adapted to machines with limited available resources. A distributed dichotomous algorithm (DDA) is proposed for association rule mining. The main features of DDA are that this algorithm does not require a high level of synchronization and that it does not process data replication and redundant calculations. In addition, DDA can partition recursively the tasks and the data set so as to be processed by machines with limited available resources.