DISCOVERING INDIRECT GENE ASSOCIATIONS BY FILTERING-BASED INDIRECT ASSOCIATION RULE MINING

摘要

Data mining is a popular technology used for microarray analysis. Using this technique, biologists can effectively elucidate gene expression data. In this research, we propose the FIARM (Filtering-Based Indirect Association Rule Mining) algorithm to analyze gene microarray data. The form (X, YM} is used to present the indirect relation of X and Y, which depends on M. This signifies that both gene X and gene M are likely involved in a given biological activity. Furthermore, both gene Y and gene M likely join together to carry out another biological activity. As gene M is the necessary factor in these different biological activities, it can help biologists determine gene relationships in diverse activities. We use semantic similarity of Gene Ontology to verify the accuracy of discovered gene relations. Under experimental evaluation, the proposed method can discover the relationship dissimilated by association rules to effectively assist biologists in complicated genetic research.