Feature Selection for Classification Using an Ant Colony System

Abd-Alsabour, Nadia; Randall, Marcus

doi:10.1109/esciencew.2010.23

Cited by 19 publications

(11 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The binary representation is also commonly used within swarm intelligence FSS algorithms. Abd‐Alsabour and Randall propose an ACO algorithm in which each ant has an associated binary vector. At the construction step, ants select one feature to be included, i , by a probability calculation.…”

Section: Fss For Classificationmentioning

confidence: 99%

Evolutionary computation for feature selection in classification problems

Iglesia

2013

WIREs Data Min & Knowl

View full text Add to dashboard Cite

Feature subset selection (FSS) has received a great deal of attention in statistics, machine learning, and data mining. Real world data analyzed by data mining algorithms can involve a large number of redundant or irrelevant features or simply too many features for a learning algorithm to handle them efficiently. Feature selection is becoming essential as databases grow in size and complexity. The selection process is expected to bring benefits in terms of better performing models, computational efficiency, and simpler more understandable models. Evolutionary computation (EC) encompasses a number of naturally inspired techniques such as genetic algorithms, genetic programming, ant colony optimization, or particle swarm optimization algorithms. Such techniques are well suited to feature selection because the representation of a feature subset is straightforward and the evaluation can also be easily accomplished through the use of wrapper or filter algorithms. Furthermore, the capability of such heuristic algorithms to efficiently search large search spaces is of great advantage to the feature selection problem. Here, we review the use of different EC paradigms for feature selection in classification problems. We discuss details of each implementation including representation, evaluation, and validation. The review enables us to uncover the best EC algorithms for FSS and to point at future research directions. © 2013 John Wiley & Sons, Ltd. How to cite this article:WIREs Data Mining Knowl Discov 2013Discov , 3:381-407. doi: 10.1002Discov /widm.1106 INTRODUCTION The problems generated by large datasets that need to be analyzed have been surfacing since the 1980s in the context of machine learning, 1-4 statistics, [5][6][7][8][9] and data mining. 10,11Structured data used for analysis generally consists of a number of features (i.e., variables, attributes, or columns) that capture the characteristics of real world objects, and a number of representative objects or observations being captured as rows or samples. Statistics, machine learning, and data mining all offer techniques that can infer models from such data. In traditional statistics, there are a few wellchosen features and a larger number of observations * Correspondence to: bli@cmp.uea.ac.uk School of Computing Sciences, University of East Anglia, Norwich, UK Conflict of interest: The authors have declared no conflicts of interest for this article. often collected purposefully for an analysis task. Machine learning focuses on developing learning models and often uses small and reasonably wellformed training data sets for the learning task. Data mining, however, is a discipline focused on developing or adapting analysis techniques for routinely collected real world data, i.e., large, noisy, and uncertain data. It is worth noting that many of the techniques used in data mining originated with in the statistics or machine learning community.Data available for the analysis may include many irrelevant or redundant features. A feature is considered to be relevan...

show abstract

Section: Fss For Classificationmentioning

confidence: 99%

Evolutionary computation for feature selection in classification problems

Iglesia

2013

WIREs Data Min & Knowl

View full text Add to dashboard Cite

show abstract

“…From these initial positions, they traverse edges probabilistically until a traversal stopping criterion is satisfied. Abd-Alsabour and Randall [48] did not use the constructive graph. Instead, they implemented a binary ant colony system with the use of support vector machine classifier.…”

Section: ) Ant Colony Optimization For Feature Selectionmentioning

confidence: 98%

A Review on Evolutionary Feature Selection

Abd-Alsabour

2014

2014 European Modelling Symposium

Self Cite

View full text Add to dashboard Cite

This paper presents a review of some of the most recent evolutionary algorithms used for solving feature selection based upon previously published research on feature selection. In addition, we discuss various research issues relating to each of the presented evolutionary algorithm. Evolutionary algorithms present several advantages over traditional search such as they require less domain-specific information. Such advantages have made them very popular within feature selection as explained in this paper. This paper covers the first part only of the evolutionary algorithms for the feature selection problem due to the limitation of the number of pages. The references cited in this paper cover the major theoretical issues, and provide access to the main branches of the literature dealing with such methods.

show abstract

“… Besides, many researchers have applied evolutionary algorithms for the solution of the feature selection problem and they all tested the effect of adding the phase of feature selection on the performance of the used predictor by performing two types of experiments: 1. the used predicator that uses the entire set of features (without the phase of feature selection), and 2. the used predicator that uses a subset of features selected by the used algorithm for performing the feature selection. An example of that is the work of Abd-Alsabour and Randall [22] where they used the support victor machine classifier with real-world and artificial datasets. Abd-Alsabour and Randall [22] also listed other results (available from literature) of different classifiers (nearest neighbor and DistAI classifiers) with the use of different evolutionary algorithms (genetic algorithms, particle swarm optimization, and chaotic binary particle swarm optimization) for performing the feature selection phase.…”

Section: The Negative Effect Of Irrelevant Features On Predictorsmentioning

confidence: 99%

“…An example of that is the work of Abd-Alsabour and Randall [22] where they used the support victor machine classifier with real-world and artificial datasets. Abd-Alsabour and Randall [22] also listed other results (available from literature) of different classifiers (nearest neighbor and DistAI classifiers) with the use of different evolutionary algorithms (genetic algorithms, particle swarm optimization, and chaotic binary particle swarm optimization) for performing the feature selection phase.…”

Section: The Negative Effect Of Irrelevant Features On Predictorsmentioning

confidence: 99%

On the Role of Dimensionality Reduction

Abd-Alsabour¹

2018

JCP

Self Cite

View full text Add to dashboard Cite

Due to the need for the dimensionality reduction in many applications such as real-world and large scale applications, this paper demonstrates it, and covers it from all of its perspectives such as its components, its importance, and where it is needed. Moreover, this paper demonstrates the impact of irrelevant features on the performance of many predictors such as decision tree, Naï ve Bayes, nearest neighbor, and support vector machines. This is because the performance of these predictors degenerates when provided with data containing irrelevant features.

show abstract

Feature Selection for Classification Using an Ant Colony System

Cited by 19 publications

References 19 publications

Evolutionary computation for feature selection in classification problems

Evolutionary computation for feature selection in classification problems

A Review on Evolutionary Feature Selection

On the Role of Dimensionality Reduction

Contact Info

Product

Resources

About