A filter-based bare-bone particle swarm optimization algorithm for unsupervised feature selection

Zhang, Yong; Li, Haigang; Wang, Qing; Peng, Chao

doi:10.1007/s10489-019-01420-9

Cited by 83 publications

(37 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the future, a multidimensional knapsack problem, quadratic knapsack problem, knapsack sharing problem, and randomized time-varying knapsack problem can be considered to investigate the performance of MBO. Thirdly, some typical combinatorial optimization problems, such as job scheduling problems [70][71][72], feature selection [73][74][75], and classification [76], deserve serious investigation and discussion. For these challenging engineering problems, the key issue is how to encode and process constraints.…”

Section: Discussionmentioning

confidence: 99%

A Novel Monarch Butterfly Optimization with Global Position Updating Operator for Large-Scale 0-1 Knapsack Problems

Feng

Wang

2019

Mathematics

View full text Add to dashboard Cite

As a significant subset of the family of discrete optimization problems, the 0-1 knapsack problem (0-1 KP) has received considerable attention among the relevant researchers. The monarch butterfly optimization (MBO) is a recent metaheuristic algorithm inspired by the migration behavior of monarch butterflies. The original MBO is proposed to solve continuous optimization problems. This paper presents a novel monarch butterfly optimization with a global position updating operator (GMBO), which can address 0-1 KP known as an NP-complete problem. The global position updating operator is incorporated to help all the monarch butterflies rapidly move towards the global best position. Moreover, a dichotomy encoding scheme is adopted to represent monarch butterflies for solving 0-1 KP. In addition, a specific two-stage repair operator is used to repair the infeasible solutions and further optimize the feasible solutions. Finally, Orthogonal Design (OD) is employed in order to find the most suitable parameters. Two sets of low-dimensional 0-1 KP instances and three kinds of 15 high-dimensional 0-1 KP instances are used to verify the ability of the proposed GMBO. An extensive comparative study of GMBO with five classical and two state-of-the-art algorithms is carried out. The experimental results clearly indicate that GMBO can achieve better solutions on almost all the 0-1 KP instances and significantly outperforms the rest.

show abstract

Section: Discussionmentioning

confidence: 99%

A Novel Monarch Butterfly Optimization with Global Position Updating Operator for Large-Scale 0-1 Knapsack Problems

Feng

Wang

2019

Mathematics

View full text Add to dashboard Cite

show abstract

“…As a wrapper-based meta-heuristic feature selection algorithm for optimal/highly discriminative feature selection, the BPSO helps minimize (if not eliminate) the curse of dimensionality caused by high-dimensional features; however, since the past decade when the canonical PSO was introduced, more than 4000 variants of the PSO algorithm has been developed including but not limited to the multiobjective PSO for cost-based feature selection in classification [39], variable-size cooperative co-evolutionary PSO for feature selection on high-dimensional data [40], Quantum PSO [41], Comprehensive Learning PSO [42], etc. Also, its capabilities as a filter-based feature selection approach has been recently explored by the authors of [43] who introduced the novel filter-based bare-bone particle swarm optimization(FBPSO) algorithm for unsupervised feature selection in cases of unlabelled data. Nevertheless, every PSO variant, though as effective as the others, is unique in architecture while targeted systems are usually Black − box optimization problems; hence, it would be futile to even attempt to assess and compare all these PSO variants for feature selection [44]; however, considering the computational cost efficiency, availability, ease-of-use, stability and robustness (It also exhibits conceptual similarities with reinforcement learning approaches [45]) associated with the BPSO, this study employs the BPSO.…”

Section: B Meta-heuristic Methods For Feature Selectionmentioning

confidence: 99%

A Multi-Domain Diagnostics Approach for Solenoid Pumps Based on Discriminative Features

Akpudo

Hur

2020

IEEE Access

View full text Add to dashboard Cite

Accurate condition monitoring of industrial cyber-physical systems/components demands the use of reliable fault detection and isolation (FD&I) methodologies. Meta-heuristic algorithms for feature selection have good exploration capability for optimal discriminative feature selection for fault isolation/classification of which the Binary Particle swarm optimization (BPSO) is superior to its counterparts. This study presents a robust approach for vibration-based failure diagnostics of electromagnetic/solenoid pumps which employ a multi-domain feature extraction procedure (statistical time-domain and frequencydomain features, Mel frequency cepstral coefficients, and continuous wavelet coefficients) for capturing linear and nonlinear properties from the signals. Compared with other filter and wrapper methods for supervised feature selection, a hybrid filter-wrapper (Pearson's correlation-BPSO (ρ-BPSO)) feature selection procedure is proposed for global search of optimal discriminative (uncorrelated) features for fault diagnosis with an RBF-kernel support vector machine (SVM*). Subsequently, a practical case study involving five VSC63A5 solenoid pumps at various operating/fault conditions is presented for validating the performance of the proposed approach. Results show the superior performance of the proposed hybrid filterwrapper approach against filter-based and wrapper-based techniques for discriminative feature selection. Also, the proposed ρ-BPSO-SVM* diagnostics model performance was compared with other standard fault isolation/classification methods.

show abstract

“…Various methods have been developed for the task of feature selection in the unsupervised setting. Most of existing works distinguish these algorithms into three groups, i.e., filter [2], [4], [10], wrapper and embedded approaches [11]- [13], in terms of different selection strategy. Moreover, with the absent of supervised information, one of the key problem for unsupervised feature selection is to design the appropriate criterion to guide the search of relevant and informative features.…”

Section: Related Work a Unsupervised Feature Selectionmentioning

confidence: 99%

Unsupervised Dual Learning for Feature and Instance Selection

Ren

Zhou

et al. 2020

IEEE Access

View full text Add to dashboard Cite

Feature selection and instance selection are dual operations on a data matrix. Feature selection aims at selecting a subset of relevant and informative features from original feature space, while instance selection at identifying a subset of informative and representative instances. Most of previous studies address these two problems separately, such that irrelevant features (resp. outliers) may mislead the process of instance (resp. feature) selection. In this paper, we address the problem by doing feature and instance selection simultaneously. We propose a novel unified framework, which chooses instances and features simultaneously, such that 1)all the data can be reconstructed from the selected instances and features and 2) the global structure which is characterized by the sparse reconstruction coefficient is preserved. Experimental results on several benchmark data sets demonstrate the effectiveness of our proposed method.

show abstract

A filter-based bare-bone particle swarm optimization algorithm for unsupervised feature selection

Cited by 83 publications

References 34 publications

A Novel Monarch Butterfly Optimization with Global Position Updating Operator for Large-Scale 0-1 Knapsack Problems

A Novel Monarch Butterfly Optimization with Global Position Updating Operator for Large-Scale 0-1 Knapsack Problems

A Multi-Domain Diagnostics Approach for Solenoid Pumps Based on Discriminative Features

Unsupervised Dual Learning for Feature and Instance Selection

Contact Info

Product

Resources

About