Evolutionary Feature Selection Method via A Chaotic Binary Dragonfly Algorithm

Feature selection aims at reducing the number of attributes while achieving a high classification accuracy in machine learning. In this paper, we design a fitness function to jointly reduce the number of the selected features and enhance the accuracy. Then, we propose a chaotic binary dragonfly algo...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics S. 2471 - 2478
Hauptverfasser: Liu, Zhao, Wang, Aimin, Sun, Geng, Li, Jiahui, Bao, Haiming, Li, Hongjuan
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 09.10.2022
Schlagworte:
chaotic map
classification
Classification algorithms
dragonfly algorithm
evolutionary population dynamics
Feature extraction
Feature selection
Heuristic algorithms
Machine learning
Machine learning algorithms
Sociology
Statistics
ISSN:2577-1655
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Feature selection aims at reducing the number of attributes while achieving a high classification accuracy in machine learning. In this paper, we design a fitness function to jointly reduce the number of the selected features and enhance the accuracy. Then, we propose a chaotic binary dragonfly algorithm (CBDA) with several improved factors on the conventional dragonfly algorithm (DA) for developing a wrapper-based feature selection method to solve the fitness function. Specifically, the CBDA introduces three improved factors that are the chaotic map, evolutionary population dynamics mechanism and binarization strategy to make the algorithm more suitable for the problem. Experiments are conducted to evaluate the performance of the proposed CBDA on 24 well-known data sets from the UCI repository, and the results demonstrate that the proposed CBDA outperforms other comparative algorithms on the majority of the tested data sets.
ISSN:2577-1655
DOI:10.1109/SMC53654.2022.9945264