Feature selection is a crucial method for discovering relevant features in high-dimensional data. However, most studies primarily focus on completely labeled data, ignoring the frequent occurrence of missing labels in real-world problems. To address high-dimensional and label-missing problems in data classification simultaneously, we proposed a semisupervised bacterial heuristic feature selection algorithm. To track the label-missing problem, a k-nearest neighbor semisupervised learning strategy is designed to reconstruct missing labels. In addition, the bacterial heuristic algorithm is improved using hierarchical population initialization, dynamic learning, and elite population evolution strategies to enhance the search capacity for various feature combinations. To verify the effectiveness of the proposed algorithm, three groups of comparison experiments based on eight datasets are employed, including two traditional feature selection methods, four bacterial heuristic feature selection algorithms, and two swarm-based heuristic feature selection algorithms. Experimental results demonstrate that the proposed algorithm has obvious advantages in terms of classification accuracy and selected feature numbers.
Bacterial Foraging Optimization (BFO) has been predominately applied to some real-world problems, but this method has poor convergence speed over complex optimization problems. In this paper, an improved Bacterial Foraging Optimization with Differential and Poisson Distribution strategies (PDBFO) is proposed to promote the insufficiency of BFO. In PDBFO, the step size of bacteria is segmented and adjusted in accordance with fitness value to accelerate convergence and enhance the search capability. Moreover, the differential operator and the Poisson Distribution strategy are incorporated to enrich individual diversity, which prevents algorithm from being trapped in the local optimum. Experimental simulations on eleven benchmark functions demonstrate that the proposed PDBFO has better convergence behavior in comparison to other six algorithms. Additionally, to verify the effectiveness of the method in solving the real-world complex problems, the PDBFO is also applied to the Nurse Scheduling Problem (NSP). Results indicate that the proposed PDBFO is more effective in obtaining the optimal solutions by comparing with other algorithms.
Efficient classification methods can improve the data quality or relevance to better optimize some Internet applications such as fast searching engine and accurate identification. However, in the big data era, difficulties and volumes of data processing increase drastically. To decrease the huge computational cost, heuristic algorithms have been used. In this paper, an Adapting Chemotaxis Bacterial Foraging Optimization (ACBFO) algorithm is proposed based on basic Bacterial Foraging Optimization (BFO) algorithm. The aim of this work is to design a modified algorithm which is more suitable for data classification. The proposed algorithm has two updating strategies and one structural changing. First, the adapting chemotaxis step updating strategy is responsible to increase the flexibility of searching. Second, the feature subsets updating strategy better combines the proposed heuristic algorithm with the KNN classifier. Third, the nesting structure of BFO has been simplified to reduce the computation complexity. The ACBFO has been compared with BFO, BFOLIW and BPSO by testing on 12 widely used benchmark datasets. The result shows that ACBFO has a good ability of solving classification problems and gets higher accuracy than the other comparation algorithm.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.