A new unsupervised fuzzy feature ranking measure for feature evaluation

Feature selection and feature ranking is a preprocessing step for data mining tasks, to reduce dimensionality, removing irrelevant data, increasing learning accuracy, and improving result comprehensibility. Filter-based feature ranking techniques rank the features according to their relevance and we choose the most relevant features to build classification models subsequently. In this paper, we propose a new unsupervised filter ranking method which uses fuzzy clustering and fuzzy entropy for ranking the features. The results are compared with three famous ranking methods. The quality of the feature subsets with highest ranks is evaluated by using five classifiers. The results obtained show that our method is effective in terms of ranking the relevant features.