Data‐driven approach to weekly forecast of the western flower thrips ( <scp> <i>Frankliniella occidentalis</i> </scp> Pergande) population in a pepper greenhouse with an ensemble model
Our results demonstrated the benefits of the ensemble model over the traditional 'single-best model' approach, avoiding model structural biases and minimizing the risk of overconfidence. This showcased how an ensemble model with minimal training data can assist growers in fully utilizing the pest monitoring data and support their decision-making on IPM. © 2025 Society of Chemical Industry.
Authors
- Herwig Leirs (ORCID: https://orcid.org/0000-0002-7612-5024)
- Nathalie Brenard (ORCID: https://orcid.org/0000-0002-9495-4985)
- Vincent Sluydts (ORCID: https://orcid.org/0000-0001-5213-851X)
- Kin Ho Chan (ORCID: https://orcid.org/0000-0003-2573-4898)
- Marlies Huysmans
- Rob Moerkens (ORCID: https://orcid.org/0000-0002-9882-9313)
Institutions
- University of Antwerp (BE)
- Teledyne FLIR (Belgium) (BE)
- Best Biotech (US)
Publication Details
- Journal
- Pest Management Science
- Published
- 2025-02-21
- DOI
- https://doi.org/10.1002/ps.8713
- Citations
- 3
- Primary Topic
- Insect-Plant Interactions and Control
- Type
- article
- Field-Weighted Citation Impact
- 5.02
Funders
- Agentschap Innoveren en Ondernemen