Hypertension prediction by multi-objective optimization methods

Feature selection is the important part of microarray analysis and it aims finding most representative subset of the bio-markers. But selection process is a challenging task due to the high dimensional nature of gene expression data. This should also be independent of sample variations in the dataset. In this paper we present a novel hybrid method that incorporates a multi-objective optimization method, called Pareto Optimal approach (PO) with Analytical Hierarchy Process (AHP). Firstly, PO was used to selects relevant subsets of the attributes, but it does not give any information about priorities of the selected bio-markers. In order to prevent this problem, AHP is incorporated with PO. AHP prioritize the selected genes by PO. This is further supported with different biomarker selection methods. The proposed method was tested on hypertension prediction.