An Extensive Empirical Study of Feature Selection Metrics for Text Classification

By George Forman. 2003

3,002 citations recorded by the source corpus.

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

Domain Adaptation for Statistical Classifiers (2006, methodology): “Specifically, we took the 10k top features according to the information gain criteria to predict “indomain” versus “out-of-domain” (as opposed to feature selection for class label); Forman (2003) provides an overview of different selection techniques.1”
A review of feature selection methods on synthetic data (2013, methodology): “So, for example, the work in [10] presents an empirical comparison of twelve feature selection methods evaluated on a benchmark of 229 text classification problem instances; the comparative study in [28] is used for the detection of breast cancers in mammograms; recent works analyze the application ”
Document-level sentiment classification: An empirical comparison between SVM and ANN (2013, methodology): “None of them has been widely accepted as the best feature selection method for sentiment classification or text categorization, however, information gain has often been competitive (Abbasi et al., 2011; Forman, 2003; Li et al., 2009; Xia & Zong, 2010; Yang & Pedersen, 1997).”
Robust Malware Detection for Internet of (Battlefield) Things Devices Using Deep Eigenspace Learning (2019, methodology): “Information retrieval techniques are widely used for feature selection [29].”
Combating the Small Sample Class Imbalance Problem Using Feature Selection (2010, methodology): “In this time period, a number of researchers have conducted research on using feature selection to combat the class imbalance problem [4], [5].”
A Review of Feature Selection Methods for Machine Learning-Based Disease Risk Prediction (2022, background): “Several studies have compared the predictive performance of the different feature selection methods (Forman, 2003; Bolón-Canedo et al., 2013; Aphinyanaphongs et al., 2014; Wah et al., 2018; Bommert et al., 2020).”
On Similarity Preserving Feature Selection (2013, methodology): “As an important technique for dimensionality reduction, feature selection has been applied to various areas, including computer vision [10], text mining [13], and bioinformatics [32], to name a few.”
Comparative study on methods of detecting research fronts using different types of citation (2009, methodology): “First, in style-based text categorization, the most significant features are the most frequent ones (Houvardas & Stamatatos, 2006; Koppel, Akiva, & Dagan, 2006) while in topicbased text categorization, the best features should be selected based on their discriminatory power (Forman, 2003).”
A review of unsupervised feature selection methods (2019, background): “However, due to the technological development raised in the last years, as well as the vast amount of unlabeled data generated in different applications such as text mining (Feldman and Sanger 2006; Bharti and kumar Singh 2014; Forman 2003), bioinformatics (Saeys et al.”
A survey of modern authorship attribution methods (2009, methodology): “First, in style-based text categorization, the most significant features are the most frequent ones (Houvardas & Stamatatos, 2006; Koppel, Akiva, & Dagan, 2006) while in topicbased text categorization, the best features should be selected based on their discriminatory power (Forman, 2003).”
Enquiring Minds: Early Detection of Rumors in Social Media from Enquiry Posts (2015, methodology): “Chi-Squared test is a classical statistical test of independence and the score measures the divergence from the expected distribution if one assumes a feature is independent of the class label [12].”
Support vector machines classification with a very large-scale taxonomy (2005, methodology): “2) We selected 4000 features for each binary classification task using the CHI algorithm [7][ 8 ].”