Skip to main content

Table 2 BPSO based feature selection methods

From: SDPSO: Spark Distributed PSO-based approach for feature selection and cancer disease prognosis

Authors Title Contributions
Chuang et al. [20] Improved binary PSO for feature selection using gene expression data An improved BPSO (IBPSO) to implement feature selection for gene expression data along with the K-nearest neighbor (KNN) method serving as an evaluator of the IBPSO as a classifier for expression data. The authors affirm a 2.85% higher accuracy compared to the previously best results published
Yang et al. [21] Boolean binary particle swarm optimization for feature selection A Boolean function which improves on the disadvantages of standard BPSO and use it to implement feature selection tasks for six microarray datasets. The experimental results also illustrate that the proposed method improves the performance on clustering gene expression data in accuracy
Behjat et al. [22] A New Binary Particle Swarm Optimization for Feature Subset Selection with Support Vector Machine A novel feature selection method called the New Binary Particle Swarm Optimization (NBPSO) to choose a set of optimal features. The proposed feature selection method was tested in classification experiments using a Support-Vector Machine (SVM) model to classify emails according to the various features as input
Wei et al. [23] A BPSO-SVM algorithm based on memory renewal and enhanced mutation mechanisms for feature selection A mutation enhancement of the BPSO-SVM algorithm through adjusting the memory of Local and Global Optimum (LGO). The algorithm also aims to increase particles’ mutation probability for feature selection to overcome convergence premature problems and achieve high quality features. Experimental results carried out on numerous datasets indicate that the proposed algorithm improved the accuracy and decreased the number of feature subsets
Kumar et al. [24] An improved BPSO algorithm for feature selection A hybrid feature selection approach BPSO–SCA. The approach performs cluster analysis by employing a cross breed technique of Binary Particle Swarm Optimization (BPSO) and Sine Cosine Algorithm (SCA) designated as Hybrid Binary Particle Swarm Optimization and Sine Cosine Algorithm (HBPSOSCA), which aims to increase the analysis accuracy