Search

In recent years, Unsupervised Feature Selection (UFS) methods have attracted considerable interest in different research areas due to their wide application in problems where unlabeled data appears. Nevertheless, few UFS methods can process datasets described by both numerical and non-numerical features (mixed data). This paper introduces a new filter UFS method based on a new feature correlation measure for mixed data to select a relevant and non-redundant feature subset. The proposed method addresses the feature selection problem in two stages through a strategy that combines Spectral Feature Selection to identify relevant features and a Pair-wise Redundancy Analysis to remove those features with a high correlation with others. According to our experiments, the proposed feature selection method based on the introduced feature correlation measure provides better quality results than previous filter UFS methods for mixed data reported in the literature. • A filter Unsupervised Feature Selection method for mixed data is presented. • A new feature correlation measure for mixed data is introduced. • Spectral Feature Selection and Pair-wise Redundancy Analysis were applied. • A comprehensive evaluation of the proposed method is provided. [ABSTRACT FROM AUTHOR]

Most of the current unsupervised feature selection methods are designed to process only numerical datasets. Therefore, in practical problems, where the objects under study are described through both numerical and non-numerical features (mixed datasets), these methods cannot be directly applied. In this work, we propose a new unsupervised filter feature selection method that can be used on datasets with both numerical and non-numerical features. The proposed method is inspired by the spectral feature selection, by using together a kernel and a new spectrum based feature evaluation measure for quantifying the feature relevance. Experiments on synthetic datasets show that in the 99% of the cases where the relevant features are known our method identifies and ranks the most relevant features at the beginning of a sorted list. Additionally, we contrast our method against state-of-the-art unsupervised filter methods over real datasets, and our method in most cases significantly outperforms them. [ABSTRACT FROM AUTHOR]

• A systematic evaluation of filter Unsupervised Feature Selection methods is presented. • The most popular and recent filter UFS methods are included in our study. • The evaluation of the filter UFS methods followed the standards in the literature. • A general discussion based on the results of the evaluated methods is provided. • Some guidelines for the use of the evaluated filter UFS methods is also provided. Unsupervised Feature Selection (UFS) has aroused great interest in the last years because of its practical significance and application on a large variety of problems in expert and intelligent systems where unlabeled data appear. Specifically, Unsupervised Feature Selection methods based on the filter approach have received more attention due to their efficiency, scalability, and simplicity. However, in the literature, there are no comprehensive studies for assessing such UFS methods when they are applied, under the same conditions, to a wide variety of real-world data. To fill this gap, in this paper, we present a comprehensive empirical and systematic evaluation of the most popular and recent filter UFS methods, evaluating their performance in terms of clustering, classification, and runtime. The filter methods used in our study were applied on 50 datasets from the UCI Machine Learning Repository and 25 high dimensional datasets from the ASU Feature Selection Repository. To evaluate if the outcomes obtained by the assessed methods are statistically significant, the Friedman test and Holm post hoc procedure were applied in the clustering and classification results. From our experiments, we provide some practical guidelines and insights for the use of the filter UFS methods analyzed in our study. [ABSTRACT FROM AUTHOR]