Equalized orthogonal map-based DNA microarray data analysis for cancer diagnosis

A new approach to cancer diagnosis involves the analysis of gene expression data. However, to make sense of multidimensional DNA microarray data is a challenging task. Beyond 3 dimensions, it is very difficult for humans to make intuitive interpretation of how each data point is distributed, how they are interrelated, and how they should be categorized. Equalized Orthogonal Map (EOM), an approach to dimension reduction was able to produce topologically correct mappings for visualization and classification of cancer types. Performance comparison with a Self-Organizing Map (SOM) is also discussed. Experimental results show that EOM performs well with DNA microarray data classification.