Approximation is widely used approach in Microarray Gene expression data analysis in terms of data vagueness or roughness, missing values and uncertainty associated with the gene expression data irrespective of it source of generation or collection. To the rescue of this problem, rough and fuzzy sets concepts based clustering started attracting world of gene expression clustering analysis. We present here an unsupervised and efficient rough-fuzzy based clustering algorithm inbuilt with validation of the generated clusters dynamically. It also carefully handles the issue of outliers in an efficient manner by not allowing them to interfere with the similarly expressed gene clusters on the fly. To achieve high end and reliable clusters, the cluster validity technique used here are Bayesian Method, Hubert’s statistics, Dunn index, β index. We vary the degree of fuzziness set for calculating the lower and upper approximation of the gene expression data. Using these validity measures, we also carried out some comparisons on the quality of clusters formed with, Fuzzy C-means (FCM), Rough C-means (RCM), Rough-Fuzzy C-means (RFCM) and Robust RFCM (rRFCM). © 2014, Research India Publications.