A hybrid Artificial Immune optimization for high-dimensional feature selection

Knowledge-Based Systems, Journal Year: 2022, Volume and Issue: 260, P. 110111 - 110111

Published: Nov. 19, 2022

Language: Английский

---

Information-Theory-based Nondominated Sorting Ant Colony Optimization for Multiobjective Feature Selection in Classification

IEEE Transactions on Cybernetics, Journal Year: 2022, Volume and Issue: 53(8), P. 5276 - 5289

Published: Aug. 22, 2022

Feature selection (FS) has received significant attention since the use of a well-selected subset features may achieve better classification performance than that full in many real-world applications. It can be considered as multiobjective optimization consisting two objectives: 1) minimizing number selected and 2) maximizing performance. Ant colony (ACO) shown its effectiveness FS due to problem-guided search operator flexible graph representation. However, there lacks an effective ACO-based approach for handle problematic characteristics originated from feature interactions highly discontinuous Pareto fronts. This article presents Information-theory-based Nondominated Sorting ACO (called INSA) solve aforementioned difficulties. First, probabilistic function is modified based on information theory identify importance features; second, new strategy designed construct solutions; third, novel pheromone updating devised ensure high diversity tradeoff solutions. INSA's compared with four machine-learning-based methods, representative single-objective evolutionary algorithms, six state-of-the-art ones 13 benchmark datasets, which consist both low high-dimensional samples. The empirical results verify INSA able obtain solutions using whose count similar or less those obtained by peers.

Language: Английский

---

Correlation-Guided Updating Strategy for Feature Selection in Classification With Surrogate-Assisted Particle Swarm Optimization

IEEE Transactions on Evolutionary Computation, Journal Year: 2021, Volume and Issue: 26(5), P. 1015 - 1029

Published: Dec. 13, 2021

Classification data are usually represented by many features, but not all of them useful. Without domain knowledge, it is challenging to determine which features Feature selection an effective preprocessing technique for enhancing the discriminating ability data, a difficult combinatorial optimization problem because challenges huge search space and complex interactions between features. Particle swarm (PSO) has been successfully applied feature due its efficiency easy implementation. However, most existing PSO-based methods still face falling into local optima. To solve this problem, article proposes novel approach, can continuously improve quality population at each iteration. Specifically, correlation-guided updating strategy based on characteristic developed, effectively use information current generate more promising solutions. In addition, particle surrogate presented, efficiently select particles with better performance in both convergence diversity form new population. Experimental comparing proposed approach few state-of-the-art 25 classification problems demonstrate that able smaller subset higher accuracy cases.

Language: Английский

---

Spatial bound whale optimization algorithm: an efficient high-dimensional feature selection approach

Neural Computing and Applications, Journal Year: 2021, Volume and Issue: 33(23), P. 16229 - 16250

Published: July 20, 2021

Language: Английский

---

Ant colony optimization equipped with an ensemble of heuristics through multi-criteria decision making: A case study in ensemble feature selection

Applied Soft Computing, Journal Year: 2022, Volume and Issue: 124, P. 109046 - 109046

Published: May 25, 2022

Language: Английский

---