Gene selection in microarray data from multi-objective perspective

Microarray technology provides a platform to study expression level of thousands of genes simultaneously, but its high dimensionality and noisy nature forces the usage of dimensionality reduction techniques. Among these techniques feature selection seems to be more favorable due to its goal to preserve feature semantic. Feature selection is also called gene selection while applied to genetic data. Inherently, gene selection objectives are manifold which makes it a proper candidate for multi-objective optimization. There are three different ways to deal with fitness evaluation in multi-objective literature. Between these three the Pareto base approach seems to deliver more promising advantages to the biologist, but it did not grab that much attention till now, probably due to its computational complexity. The intention of this paper is to provide an insight to gene selection problem from multi-objective perspective. Although, covering all the proposed methods are impossible, but hopefully those algorithms discussed here are enough to show the common trend in multi-objective gene selection in microarray data.