Most of the real-world data mining applications are characterized by high dimensional data, where not all of the features are important. High dimensional data can contain a lot of irrelevant and noisy information that may greatly degrade the performance of a data mining process. Feature selection methods are the techniques that select a subset of relevant feature for building robust learning models by removing most irrelevant and redundant features from the data. Many feature selection methods have been developed to reduce the dimensionality of big data. Among them, the Relief algorithm is general and successful attribute estimator. The main idea of Relief algorithm is to compute ranking scores for every feature indicating how well this feature separates neighboring samples. In this study, we do perform the sensitivity analysis to find the optimal number of features and also suggest the two-stage method to design the optimal feature subset.