Parallel Mining of Association Rules from Gene Expression Databases

Gene expression (GE) databases are high-dimensional and have a large number of genes/columns in each profile experiment. Recently, mining association rules (ARM) from these databases has attracted much interest. Several sequential ARM methods have been devised for this purpose, but, to the best of our knowledge, there is no parallel ARM specially designed for GE databases. On the other hand, the existing parallel ARMs are not suitable for such databases because they usually do not take into account the high-dimensionality of the data. In this paper, we propose a parallel ARM specially designed for GE databases. We show that our method parallelizes not only the tasks of reading from databases and computing the supports, but also the tasks of computing the combination between items and generating association rules, which here are very time-consuming. We analyze computation and communication costs, speed-up, and present some experimental results on real databases, as well.