The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
La sélection de fonctionnalités basée sur l’optimisation des essaims de particules est souvent utilisée pour améliorer les performances des algorithmes d’intelligence artificielle. Cependant, son interprétabilité manque de recherches concrètes. Améliorer la stabilité de la méthode de sélection des fonctionnalités est un moyen d’améliorer efficacement son interprétabilité. Une nouvelle approche de sélection de fonctionnalités appelée Interpretable Particle Swarm Optimization est développée dans cet article. Il utilise quatre méthodes de perturbation des données et trois méthodes de sélection de caractéristiques de filtrage pour obtenir des sous-ensembles de caractéristiques stables, et adopte la carte de Fuch pour les convertir en particules initiales. En outre, il utilise une stratégie de mutation par similarité, qui applique la distance de Tanimoto pour choisir le tiers des individus les plus proches des particules précédentes pour mettre en œuvre la mutation. Onze algorithmes représentatifs et quatre ensembles de données typiques sont utilisés pour effectuer une comparaison complète avec notre approche proposée. Les indicateurs d'exactitude, de F1, de précision et de taux de rappel sont utilisés comme mesures de classification, et l'extension de l'indicateur Kuncheva est utilisée comme mesure de stabilité. Les expériences montrent que notre méthode a une meilleure interprétabilité que les algorithmes évolutionnaires comparés. De plus, les résultats des mesures de classification démontrent que l'approche proposée présente d'excellentes performances de classification globale.
Yi LIU
Defense Innovation Institute
Wei QIN
Defense Innovation Institute
Qibin ZHENG
Academy of Military Science
Gensong LI
Defense Innovation Institute
Mengmeng LI
Defense Innovation Institute
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copier
Yi LIU, Wei QIN, Qibin ZHENG, Gensong LI, Mengmeng LI, "An Interpretable Feature Selection Based on Particle Swarm Optimization" in IEICE TRANSACTIONS on Information,
vol. E105-D, no. 8, pp. 1495-1500, August 2022, doi: 10.1587/transinf.2021EDL8095.
Abstract: Feature selection based on particle swarm optimization is often employed for promoting the performance of artificial intelligence algorithms. However, its interpretability has been lacking of concrete research. Improving the stability of the feature selection method is a way to effectively improve its interpretability. A novel feature selection approach named Interpretable Particle Swarm Optimization is developed in this paper. It uses four data perturbation ways and three filter feature selection methods to obtain stable feature subsets, and adopts Fuch map to convert them to initial particles. Besides, it employs similarity mutation strategy, which applies Tanimoto distance to choose the nearest 1/3 individuals to the previous particles to implement mutation. Eleven representative algorithms and four typical datasets are taken to make a comprehensive comparison with our proposed approach. Accuracy, F1, precision and recall rate indicators are used as classification measures, and extension of Kuncheva indicator is employed as the stability measure. Experiments show that our method has a better interpretability than the compared evolutionary algorithms. Furthermore, the results of classification measures demonstrate that the proposed approach has an excellent comprehensive classification performance.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2021EDL8095/_p
Copier
@ARTICLE{e105-d_8_1495,
author={Yi LIU, Wei QIN, Qibin ZHENG, Gensong LI, Mengmeng LI, },
journal={IEICE TRANSACTIONS on Information},
title={An Interpretable Feature Selection Based on Particle Swarm Optimization},
year={2022},
volume={E105-D},
number={8},
pages={1495-1500},
abstract={Feature selection based on particle swarm optimization is often employed for promoting the performance of artificial intelligence algorithms. However, its interpretability has been lacking of concrete research. Improving the stability of the feature selection method is a way to effectively improve its interpretability. A novel feature selection approach named Interpretable Particle Swarm Optimization is developed in this paper. It uses four data perturbation ways and three filter feature selection methods to obtain stable feature subsets, and adopts Fuch map to convert them to initial particles. Besides, it employs similarity mutation strategy, which applies Tanimoto distance to choose the nearest 1/3 individuals to the previous particles to implement mutation. Eleven representative algorithms and four typical datasets are taken to make a comprehensive comparison with our proposed approach. Accuracy, F1, precision and recall rate indicators are used as classification measures, and extension of Kuncheva indicator is employed as the stability measure. Experiments show that our method has a better interpretability than the compared evolutionary algorithms. Furthermore, the results of classification measures demonstrate that the proposed approach has an excellent comprehensive classification performance.},
keywords={},
doi={10.1587/transinf.2021EDL8095},
ISSN={1745-1361},
month={August},}
Copier
TY - JOUR
TI - An Interpretable Feature Selection Based on Particle Swarm Optimization
T2 - IEICE TRANSACTIONS on Information
SP - 1495
EP - 1500
AU - Yi LIU
AU - Wei QIN
AU - Qibin ZHENG
AU - Gensong LI
AU - Mengmeng LI
PY - 2022
DO - 10.1587/transinf.2021EDL8095
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E105-D
IS - 8
JA - IEICE TRANSACTIONS on Information
Y1 - August 2022
AB - Feature selection based on particle swarm optimization is often employed for promoting the performance of artificial intelligence algorithms. However, its interpretability has been lacking of concrete research. Improving the stability of the feature selection method is a way to effectively improve its interpretability. A novel feature selection approach named Interpretable Particle Swarm Optimization is developed in this paper. It uses four data perturbation ways and three filter feature selection methods to obtain stable feature subsets, and adopts Fuch map to convert them to initial particles. Besides, it employs similarity mutation strategy, which applies Tanimoto distance to choose the nearest 1/3 individuals to the previous particles to implement mutation. Eleven representative algorithms and four typical datasets are taken to make a comprehensive comparison with our proposed approach. Accuracy, F1, precision and recall rate indicators are used as classification measures, and extension of Kuncheva indicator is employed as the stability measure. Experiments show that our method has a better interpretability than the compared evolutionary algorithms. Furthermore, the results of classification measures demonstrate that the proposed approach has an excellent comprehensive classification performance.
ER -