Dépôt numérique
RECHERCHER

Advanced Efficient Feature Selection Integrating Augmented Extreme Learning Machine and Particle Swarm Optimization for Predicting Nitrogen Use Efficiency and Yield in Corn.

Téléchargements

Téléchargements par mois depuis la dernière année

Bontemps, Josselin ORCID logoORCID: https://orcid.org/0009-0008-1283-1270; Ebtehaj, Isa ORCID logoORCID: https://orcid.org/0000-0002-6906-629X; Deslauriers, Gabriel; Rousseau, Alain N. ORCID logoORCID: https://orcid.org/0000-0002-3439-2124; Bonakdari, Hossein ORCID logoORCID: https://orcid.org/0000-0001-6169-3654 et Dessureault-Rompré, Jacynthe ORCID logoORCID: https://orcid.org/0000-0002-2812-0691 (2025). Advanced Efficient Feature Selection Integrating Augmented Extreme Learning Machine and Particle Swarm Optimization for Predicting Nitrogen Use Efficiency and Yield in Corn. Agronomy , vol. 15 , nº 1. p. 244. DOI: 10.3390/agronomy15010244.

[thumbnail of P4615_JBontemps_2025.pdf]
Prévisualisation
PDF - Version publiée
Télécharger (5MB) | Prévisualisation

Résumé

Efficient nitrogen management is crucial for improving corn productivity while minimizing environmental impacts. This study evaluates the response of corn to nitrogen fertilization using three key metrics: yield; nitrogen harvest index (NHI); and agronomic nitrogen use efficiency (ANUE). This experiment was conducted over three years (2021–2023) across 84 sites in Quebec, Canada, with five nitrogen treatments applied post-emergence (0, 50, 100, 150, 200 kg N/ha) and initial nitrogen applied at seeding (30 to 60 kg/ha). In addition, various soil health indicators, including physical, chemical, and biochemical properties, were monitored to understand their interaction with nitrogen use efficiency. Machine learning techniques, such as augmented extreme learning machine (AELM) and particle swarm optimization (PSO), were employed to optimize nitrogen recommendations by identifying the most relevant features for predicting yield and nitrogen use efficiency (NUE). The results highlight that integrating soil health indicators such as enzyme activities (β-glucosidase [BG] and N-acetyl-β-D-glucosaminidase [NAG]) and soil proteins into nitrogen management models improves prediction accuracy, leading to enhanced productivity and environmental sustainability. These findings suggest that advanced data-driven approaches can significantly contribute to more precise and sustainable nitrogen fertilization strategies.

Type de document: Article
Mots-clés libres: augmented extreme learning machine (AELM); feature selection; particle swarm optimization (PSO); nitrogen use efficiency (NUE); soil health indicators; corn yield
Centre: Centre Eau Terre Environnement
Date de dépôt: 07 mars 2025 18:48
Dernière modification: 07 mars 2025 18:48
URI: https://espace.inrs.ca/id/eprint/16305

Gestion Actions (Identification requise)

Modifier la notice Modifier la notice