Efficient Feature Selection via Analysis of Relevance and Redundancy

By Lei Yu, Huan Liu. Journal of machine learning research, 2004

2,308 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/de3b02b459a37cfc71f9675647eba8d71aa57476

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Supervised Machine Learning: A Review of Classification Techniques (2007, unknown): “A key issue in many applications is to determine which of the available input features should be used in modeling via feature selection (Yu & Liu, 2004), because it could improve the classification accuracy and scale down the required classification time.”
Proposal of a Sustainable Circular Index for Manufacturing Companies (2017, background): “Attending to this criterion, the indicators with the highest correlation should be excluded from the index construction process in order to minimize their redundancy [120].”
Feature Selection for Clustering: A Review (2018, methodology): “Relief [59], Fisher score [16], CFS [24], and FCBF [76] are among the most representative algorithms of the filter model.”
Generalized BackPropagation, Étude De Cas: Orthogonality (2016, background): “…but not limited to dimensionality reduction Turk and Pentland (1991); Hyvarinen (1999); Cai et al. (2006); Cunningham and Ghahra-mani (2015), clustering Niu et al. (2010); Kumar et al. (2011), feature selection Yu and Liu (2004) and dictionary learning Lesage et al. (2005); Vidal et al. (2005).”
Face Recognition Under Varying Illumination Using Gradientfaces (2009, methodology): “The database contains 660 images of 132 individuals, each with ﬁve images taken under outdoor lighting conditions, a part of images come from Caltech [34] and FRGC Version 2 [35] face databases.”
mr2PSO: A maximum relevance minimum redundancy feature selection method based on swarm intelligence for support vector machine classification (2011, background): “Yu and Liu [59] categorize the feature space into three classes; strongly relevant, weakly relevant and irrelevant features.”
Supersparse Linear Integer Models for Interpretable Classification (2013, background): “Some feature selection algorithms rely on analysis of relevance and redundancy [78], which could yield a more interpretable feature set.”
Recent advances in feature selection and its applications (2017, methodology): “For example, in [65], the well-known algorithm Fast Correlation-Based Filter (FCBF) [24,25] is extended to handle multi-label data through graphical model to represent the relationships among labels and features.”
A Robust-Equitable Measure for Feature Ranking and Selection (2017, methodology): “Feature selection algorithms can be categorized based on how the learning algorithms is incorporated into the selection algorithm: filter, wrapper, or embedded methods (Kohavi and John, 1997; Guyon and Elisseeff, 2003; Yu and Liu, 2004).”
Neighborhood rough set based heterogeneous feature subset selection (2008, background): “With a limited amount of training data, an excessive amount of features may cause a significant slowdown in the learning process, and may increase the risk of the learned classifier to over-fit the training data because irrelevant or redundant features confuse learning algorithms [39,40].”
Combating the Small Sample Class Imbalance Problem Using Feature Selection (2010, background): “Feature selection as a general part of machine learning and data mining algorithms has been thoroughly re-searched [2], [39], [41], [42], [43], [44], [45], but its importance to resolving the class imbalance problem is a recent development with most research appearing in the previous seven years.”
Efficient Spectral Feature Selection with Minimum Redundancy (2010, background): “It is also shown empirically that removing redundant features can result in signiﬁcant performance improvement (Hall 1999; Ding and Peng 2003; Gabrilovich and Markovitch 2004; Yu and Liu 2004; Appice et al. 2004; Duangsoithong 2009).”
Incremental wrapper-based gene selection from microarray data for cancer classification (2006, unknown): “One, denoted by SFW , uses a target learning algorithm to estimate the worth of gene subsets; the other two are subset search algorithms which exploit SF search and use correlation measures (variation of CFS, correlation-based feature selection algorithm [23]) or consistency measure (variation of FO”
Feature selection in machine learning: A new perspective (2018, background): “The features in an original set can be divided into four groups [79]: (a) completely irrelevant and noisy features, (b) weakly relevant and redundant features, (c) weakly relevant and non-redundant features, and (d) strongly relevant features.”
A review of feature selection methods on synthetic data (2013, methodology): “Besides this classification, feature selection methods can also be divided according to two approaches: individual evaluation and subset evaluation [3].”