Fast Binary Feature Selection with Conditional Mutual Information

By Franccois Fleuret. Journal of machine learning research, 2004

1,183 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/c7ae0d607d21f2f07cf761b00c11cd88572c571d

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

Variational Information Maximization for Feature Selection (2016, methodology): “Since mutual information (MI) is a general measure of dependence with several unique properties [6], many MI-based scoring functions have been proposed as filter methods [7, 8, 9, 10, 11, 12]; see [5] for an exhaustive list.”
Differential evolution for filter feature selection based on information theory and feature ranking (2018, unknown): “Due to its suitability, now it has been used in a variety of ﬁelds, including natural language processing, cryptography, pattern recognition and data analysis [31] .”
Feature selection in machine learning: A new perspective (2018, methodology): “Feature selection methods based on conditional mutual information have attracted significant attention [43, 44, 83-85].”
Evaluation Measures for Multi-class Subgroup Discovery (2009, methodology): “MI as used in [18] also differs from our approach in that it selects features that maximise the MI with respect to a single class but not all the classes.”
Combating the Small Sample Class Imbalance Problem Using Feature Selection (2010, methodology): “Koller and Sahami [57] showed that the optimal feature ranking can be calculated using Markov blankets in a backward substitution algorithm; Fleuret [58] and Peng et al. [59] derived two approximations to this optimal ranking.”
Learning Hybrid Image Templates (HIT) by Information Projection (2012, background): “In recent years, by combining different types of features in a discriminative framework [25], [26], [27], [28], [29], [30], [31], [32], much progress has been made toward improving the accuracy of object categorization.”
Estimating Mutual Information for Discrete-Continuous Mixtures (2017, methodology): “Therefore, mutual information is widely used in machine learning for canonical tasks as classification [35], clustering [32, 49, 8, 29] and feature selection [2, 13].”
Conditional Likelihood Maximisation: A Unifying Framework for Information Theoretic Feature Selection (2012, background): “This term quantifies the uncertainty still remaining in the label even when we know all possible features; it is an irreducible constant, independent of all parameters, and in fact forms a bound on the Bayes error (Fano, 1961).”
Demystifying fixed k-nearest neighbor information estimators (2016, background): “Despite its widespread use (part of scientific software packages), theoretical properties of this estimator have been largely unexplored.”
Information-Theoretic Feature Selection in Microarray Data Using Variable Complementarity (2008, methodology): “Redundancy and relevance notions have been used in several filter algorithms like Ranking [16], Markov blanket filter [23], Fast Correlation Based Filter (FCBF) [46], Relevance Filter [3], [4], Minimum Redundancy Maximum Relevance (MRMR) [33] and the Conditional Mutual Information Maximization (CMIM”
A review on Gabor wavelets for face recognition (2006, background): “Shen and Bai [88] also applied mutual information theory [ 89 ] to solve the problem and the results show that the features selected by mutual information are more discriminative due to the removal of redundancy.”
A Fast Clustering-Based Feature Subset Selection Algorithm for High-Dimensional Data (2013, background): “Both this algorithm are exactly reverse of each other.”
Heart Disease Identification Method Using Machine Learning Classification in E-Healthcare (2020, methodology): “In order to tackle the feature selection problem, we proposed Fast conditional mutual information (FCMIM) feature selection algorithm [37] in this study.”
Estimation of Information Theoretic Measures for Continuous Random Variables (2008, background): “Mutual information is extensively used to assess whether two random variables are independent [2] and has been proposed to solve the all-relevant feature selection problem [8, 24].”
The Nearest Neighbor Information Estimator is Adaptively Near Minimax Rate-Optimal (2017, unknown): “They have many applications in modern machine learning tasks, such as classification [48], clustering [46, 58, 10, 41] and feature selection [1, 17].”