MR-Apriori count distribution algorithm for parallel Action Rules discovery

Data mining deals with the extraction of hidden predictive information from large databases. One of the central tasks associated with data mining is to discover profitable actions from the dataset for the decision maker. Discovering these actions can be accomplished through extracting Action Rules from the data, which has become an attractive research topic in data mining. Several methods have been developed for the discovery of Action Rules and variety of methods for association rules in the past few years. However, with the explosive recent growth of the amounts of data, there is a need for the development of scalable methods for Action Rules discovery to accommodate the massive datasets. We are not aware of any such methods existing at this time. In this paper, we propose a novel approach - a parallel Action Rule discovery algorithm based on MapReduce paradigm through count distribution. We use Hadoop, as a scalable and distributed framework for implementing this method. Experiment shows much faster computational time for Action Rules discovery in a distributed environment compared to the traditional single machine method.