A Velocity-Guided Grey Wolf Optimization Algorithm With Adaptive Weights and Laplace Operators for Feature Selection in Data Classification

IEEE Access, Год журнала: 2024, Номер 12, С. 39887 - 39901

Опубликована: Янв. 1, 2024

The rapid growth of data quantity directly leads to the increasing feature dimension, which challenges machine learning and mining. Wrapper-based intelligent swarm algorithms are effective solution techniques. Grey Wolf Optimization (GWO) algorithm is a novel population algorithm. Simple principles few parameters characterize it. However, basic GWO has disadvantages, such as difficulty coordinating exploration exploitation capabilities premature convergence. As result, fails identify many irrelevant redundant features. To improve performance algorithm, this paper proposes velocity-guided grey wolf optimization with adaptive weights Laplace operators (VGWO-AWLO). Firstly, by introducing uniformly distributed dynamic weighting mechanism, control $a$ guided undergo nonlinear changes achieve good transition from exploratory phase development phase. Second, velocity-based position update formula designed an individual memory function enhance local search capability wolves drive them converge optimal solution. Thirdly, cross-operator strategy applied increase diversity help escape Finally, VGWO-AWLO evaluated for its comprehensive in terms classification accuracy, dimensionality approximation, convergence, stability 18 classified datasets. experimental results show that accuracy convergence speed better than GWO, variants, other state-of-the-art meta-heuristic algorithms.

Язык: Английский

---

Multi-Strategy Assisted Multi-Objective Whale Optimization Algorithm for Feature Selection

Computer Modeling in Engineering & Sciences, Год журнала: 2024, Номер 140(2), С. 1563 - 1593

Опубликована: Янв. 1, 2024

In classification problems, datasets often contain a large amount of features, but not all them are relevant for accurate classification.In fact, irrelevant features may even hinder accuracy.Feature selection aims to alleviate this issue by minimizing the number in subset while simultaneously error rate.Single-objective optimization approaches employ an evaluation function designed as aggregate with parameter, results obtained depend on value parameter.To eliminate parameter's influence, problem can be reformulated multi-objective problem.The Whale Optimization Algorithm (WOA) is widely used problems because its simplicity and easy implementation.In paper, we propose multi-strategy assisted WOA (MSMOWOA) address feature selection.To enhance algorithm's search ability, integrate multiple strategies such Levy flight, Grey Wolf Optimizer, adaptive mutation into it.Additionally, utilize external repository store non-dominant solution sets grid technology maintain diversity.Results fourteen University California Irvine (UCI) demonstrate that our proposed method effectively removes redundant improves performance.The source code accessed from website: https://github.com/zc0315/MSMOWOA.

Язык: Английский

---

A Novel Hybrid Algorithm Based on Beluga Whale Optimization and Harris Hawks Optimization for Optimizing Multi-Reservoir Operation

Water Resources Management, Год журнала: 2024, Номер 38(12), С. 4883 - 4909

Опубликована: Июнь 19, 2024

Язык: Английский

---

Targeting wastewater quality variables prediction: Improving sparrow search algorithm towards optimizing echo state network

Journal of Water Process Engineering, Год журнала: 2024, Номер 65, С. 105717 - 105717

Опубликована: Июль 13, 2024

Язык: Английский

---

A hybrid snake optimizer with crisscross learning strategy for constrained structural optimization

Engineering Optimization, Год журнала: 2025, Номер unknown, С. 1 - 36

Опубликована: Апрель 4, 2025

Язык: Английский

---