Fuzzy-Rough Feature Selection using f-Information Measures
Feature selection or dimensionality reduction of a data set is an essential preprocessing step used for pattern recognition, data mining, and machine learning. It is an important problem related to mining large data sets, both in dimension and size. Before the analysis of the data set, preprocessing the data to obtain a smaller set of representative features and retaining the optimal salient characteristics of the data not only decrease the processing time but also lead to more compactness of the models learned and better generalization. Hence, the general criterion for reducing the dimension is to preserve the most relevant information of the original data according to some optimality criteria.
Conventional methods of feature selection involve evaluating different feature subsets using some indices and selecting the best among them. Depending on the way of computing the feature evaluation index, feature selection methods are generally divided into two broad categories: filter approach [1, 2] and wrapper approach [1, 3]. In the filter approach, the algorithms do not perform classification of the data in the process of feature evaluation. Before application of the actual learning algorithm, the best subset of features is selected in one pass by evaluating some predefined criteria, which are independent of the actual generalization performance of the learning machine. Hence, the filter approach ...