Two-Step Particle Swarm Optimization to Solve the Feature Selection Problem

Bello, Rafael; Gomez, Yudel; Nowé, Ann; Garcia, Maria M.

doi:10.1109/isda.2007.4389688

Cited by 35 publications

(45 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In the second stage, only the top-ranked features are used as the candidate features for the wrapper algorithm. In [28] a PSO-based feature subset selection algorithm is proposed for the classification of high-dimensional cancer microarray data. In the first stage, the dataset is clustered by the k-means algorithm, then a filter algorithm is applied to rank each gene in every cluster.…”

Section: Literature Reviewmentioning

confidence: 99%

Ensemble of Filter-Based Rankers to Guide an Epsilon-Greedy Swarm Optimizer for High-Dimensional Feature Subset Selection

2017

View full text Add to dashboard Cite

The main purpose of feature subset selection is to remove irrelevant and redundant features from data, so that learning algorithms can be trained by a subset of relevant features. So far, many algorithms have been developed for the feature subset selection, and most of these algorithms suffer from two major problems in solving high-dimensional datasets: First, some of these algorithms search in a high-dimensional feature space without any domain knowledge about the feature importance. Second, most of these algorithms are originally designed for continuous optimization problems, but feature selection is a binary optimization problem. To overcome the mentioned weaknesses, we propose a novel hybrid filter-wrapper algorithm, called Ensemble of Filter-based Rankers to guide an Epsilon-greedy Swarm Optimizer (EFR-ESO), for solving high-dimensional feature subset selection. The Epsilon-greedy Swarm Optimizer (ESO) is a novel binary swarm intelligence algorithm introduced in this paper as a novel wrapper. In the proposed EFR-ESO, we extract the knowledge about the feature importance by the ensemble of filter-based rankers and then use this knowledge to weight the feature probabilities in the ESO. Experiments on 14 high-dimensional datasets indicate that the proposed algorithm has excellent performance in terms of both the error rate of the classification and minimizing the number of features.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Ensemble of Filter-Based Rankers to Guide an Epsilon-Greedy Swarm Optimizer for High-Dimensional Feature Subset Selection

2017

View full text Add to dashboard Cite

show abstract

“…The evaluation function of the mesh nodes is displayed in expression (2) and is the same used in [15] and [4]. It considers the number of attributes included in the feature subset R represented by node n and the quality of the classification associated to R. The goal is to maximize the function below.…”

Section: A Reduct Is a Minimal Subset Of Features B ⊆ A Such That In mentioning

confidence: 99%

“…The dynamic nature of our proposal manifests in the generation of (i) the initial mesh; During the mesh expansion in each cycle, a weight w is defined using expression (3) as in [14], [15] and [4].…”

Section: Generation Of Nodes In Dmomentioning

confidence: 99%

“…Some of them have exhibited good results, mainly attained by using ACO-or PSO-based approaches, such as [8], [1], [2], [13], [14] and [15]. In [3] and [4], a new approach to feature selection based on the ACO and PSO methodologies are presented. The chief thought is the split of the search process accomplished by the agents (ants or particles) into two stages, such that ants are commanded in the first stage to find partial solutions to the problem, which in turn are afterwards used as initial states during the upcoming phase.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Feature Selection through Dynamic Mesh Optimization

Bello

Puris

Falcón

et al. 2008

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

This paper introduces the Dynamic Mesh Optimization meta-heuristic, which falls under the evolutionary computation techniques. Moreover, we outline its application to the feature selection problem. A set of nodes representing subsets of features makes up a mesh which dynamically grows and moves across the search space. The novel methodology is compared with other existing meta-heuristic approaches, thus leading to encouraging empirical results.

show abstract

“…Another feature selection approach that based on Scatter Search (SSAR) is proposed by Jue et al in [18]. The Particle Swarm Optimization algorithm (PSO), proposed by Kennedy and Eberhart [19], was used in feature selection approaches as in [20][21][22][23]. Based on the biological behaviour of bees, Karaboga [24] proposed an optimisation approach called the Artificial Bee Colony (ABC) algorithm.…”

Section: Introductionmentioning

confidence: 99%

Binary Dragonfly Algorithm for Feature Selection

Mafarja

Eleyan

Jaber

et al. 2017

2017 International Conference on New Trends in Computing Sciences (ICTCS)

171

View full text Add to dashboard Cite

Wrapper feature selection methods aim to reduce the number of features from the original feature set to and improve the classification accuracy simultaneously. In this paper, a wrapper-feature selection algorithm based on the binary dragonfly algorithm is proposed. Dragonfly algorithm is a recent swarm intelligence algorithm that mimics the behavior of the dragonflies. Eighteen UCI datasets are used to evaluate the performance of the proposed approach. The results of the proposed method are compared with those of Particle Swarm Optimization (PSO), Genetic Algorithms (GAs) in terms of classification accuracy and number of selected attributes. The results show the ability of Binary Dragonfly Algorithm (BDA) in searching the feature space and selecting the most informative features for classification tasks.

show abstract

Two-Step Particle Swarm Optimization to Solve the Feature Selection Problem

Cited by 35 publications

References 23 publications

Ensemble of Filter-Based Rankers to Guide an Epsilon-Greedy Swarm Optimizer for High-Dimensional Feature Subset Selection

Ensemble of Filter-Based Rankers to Guide an Epsilon-Greedy Swarm Optimizer for High-Dimensional Feature Subset Selection

Feature Selection through Dynamic Mesh Optimization

Binary Dragonfly Algorithm for Feature Selection

Contact Info

Product

Resources

About