Dimension Reduction of Microarray Data Using Gene Ontology and Correlation Filter

Analysis of every variable at a microscopic level is not a feasible way. It might take a long time to perform any meaningful analysis. So, the valuable time and money are wasted for analysis of high dimensional data. In this paper, a better way is given to deal with high dimensional data and proposed a novel dimension reduction technique based on gene ontology and Pearson's correlation coefficient. The names of genes are extracted from microarray data and classify them based on biological processes of gene ontology and find their relation with cellular components. Next, the gene correlation factor is identified on every network of each of the group. Lastly, the independent component (IC) value of the genes related to the gene ontology is calculated and selects only those genes having the highest IC value of their corresponding network and eliminates the rest of the genes. It reduces size of the dataset at least 40% from its original size.