Mining high-correlation association rules for inferring gene regulation networks

Construction gene regulation networks can provide insights into the understanding the molecular mechanisms underlying important biological processes. We present a novel association rule mining for building large-scale gene regulation networks from microarray data. Gene expression microarray data typically contains a very high gene dimension and a very low sample size, rendering a great challenge for existing association rule mining algorithms. In this paper, we develop a novel algorithm, HCMiner, to mine high-correlation association rules from microarray data. HCMiner initially overlapping partitions the dimension of genes according to their correlations and introduces the support-free framework for mining association rules. Several experiments on Yeast dataset show that the proposed algorithm outperforms existing algorithms with respect to scalability and effectiveness.