Empirical study of feature selection methods based on individual feature evaluation for classification problems

Araúzo-Azofra, Antonio; Aznarte, José Luis; Benítez, José Manuel

doi:10.1016/j.eswa.2010.12.160

Cited by 92 publications

(48 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Various studies have addressed text classification using different techniques to classify text documents, and different metrics to evaluate the accuracies of these techniques [10], [12], [13], [14]. At present most of the studies are based on feature selection, and are mainly focused on its different methods performance in statistical learning on text classification [11], [15], [16], [17].…”

Section: Related Workmentioning

confidence: 99%

Empirical Study on Filter based Feature Selection Methods for Text Classification

DeySarakar¹,

Goswami²

2013

IJCA

View full text Add to dashboard Cite

Text classification has become much more relevant with the increased volume of unstructured data from various sources. Several techniques have been developed for text classification. High dimensionality of feature space is one of the established problems in text classification. Feature selection is one of the techniques to reduce dimensionality. Feature selection helps in increasing classifier performance, reduce over filtering to speed up the classification model construction and testing and make models more interpretable. This paper presents an empirical study comparing performance of few feature selection techniques (Chi-squared, Information Gain, Mutual Information and Symmetrical Uncertainty) employed with different classifiers like naive bayes, SVM, decision tree and k-NN. Motivation of the paper is to present results of feature selection methods on various classifiers on text datasets. The study further allows comparing the relative performance of the classifiers and the methods.

show abstract

Section: Related Workmentioning

confidence: 99%

Empirical Study on Filter based Feature Selection Methods for Text Classification

DeySarakar¹,

Goswami²

2013

IJCA

View full text Add to dashboard Cite

show abstract

“…A second major issue is the analysis of high-dimensional data and feature selection [4], [5], [15] which has been extensively explored in a healthcare context [1], [3], [11]. In [11] and [1], both groups describe a methodology where features are selected in a two-step manually intensive fashion in order to learn predictive models.…”

Section: Related Researchmentioning

confidence: 99%

“…In [11] and [1], both groups describe a methodology where features are selected in a two-step manually intensive fashion in order to learn predictive models. In these two approaches for selecting a feature representation in the health domain, shallow algorithms are utilised and high dimensional data is not encountered, where in one instance only nine features were modelled [11].…”

Section: Related Researchmentioning

confidence: 99%

“…In these two approaches for selecting a feature representation in the health domain, shallow algorithms are utilised and high dimensional data is not encountered, where in one instance only nine features were modelled [11]. Furthermore, sometimes relevant features were completely eliminated which impacted on the performance of the model [1].…”

Section: Related Researchmentioning

confidence: 99%

See 1 more Smart Citation

A Framework for Selecting Deep Learning Hyper-parameters

Donoghue

Roantree

2015

Data Science

View full text Add to dashboard Cite

Abstract. Recent research has found that deep learning architectures show significant improvements over traditional shallow algorithms when mining high dimensional datasets. When the choice of algorithm employed, hyper-parameter setting, number of hidden layers and nodes within a layer are combined, the identification of an optimal configuration can be a lengthy process. Our work provides a framework for building deep learning architectures via a stepwise approach, together with an evaluation methodology to quickly identify poorly performing architectural configurations. Using a dataset with high dimensionality, we illustrate how different architectures perform and how one algorithm configuration can provide input for fine-tuning more complex models.

show abstract

“…The selection of features and the removal or reduction of redundant information unrelated to the classifica-tion task on hand will not only reduce the complexity of the prob-lem and improve the efficiency of the processing but will also simplify significantly the design of the classifier. The FS is one of the essential and frequently used techniques in machine learning (Arauzo-Azofra, Aznarte, & Benítez, 2010;Foithong, Pinngern, & Attachoo, 2011;García-López, García-Torres, Melián-Batista, Moreno-Pérez, & Moreno-Vega, 2006;García-Torres, García-López, Melián-Batista, Moreno-Pérez, & Moreno-Vega, 2004;Kabir, Shahjahan, & Murase, 2011;Pacheco, Casado, & Núnez, 2007;Yang, Liao, Meng, & Lee, 2011). An FS method generates different candidates from the feature space and assesses them based on an evaluation criterion to find the best feature subset (Dash & Liu, 1997).…”

Section: Introductionmentioning

confidence: 99%

Fast feature selection aimed at high-dimensional data via hybrid-sequential-ranked searches

Ruíz

Riquelme

Aguilar–Ruiz

et al. 2012

Expert Systems with Applications

View full text Add to dashboard Cite

We address the feature subset selection problem for classification tasks. We examine the performance of two hybrid strategies that directly search on a ranked list of features and compare them with two widely used algorithms, the fast correlation based filter (FCBF) and sequential forward selection (SFS). The pro-posed hybrid approaches provide the possibility of efficiently applying any subset evaluator, with a wrap-per model included, to large and high-dimensional domains. The experiments performed show that our two strategies are competitive and can select a small subset of features without degrading the classifica-tion error or the advantages of the strategies under study.

show abstract

Empirical study of feature selection methods based on individual feature evaluation for classification problems

Cited by 92 publications

References 10 publications

Empirical Study on Filter based Feature Selection Methods for Text Classification

Empirical Study on Filter based Feature Selection Methods for Text Classification

A Framework for Selecting Deep Learning Hyper-parameters

Fast feature selection aimed at high-dimensional data via hybrid-sequential-ranked searches

Contact Info

Product

Resources

About