Guiding wine-grape variety selection using AI recommender systems
Canadian Biosystems Engineering · 31.12.2025 · 67 (1) · 701
DOI: 10.7451/cbe.2025.67.7.1 ↗Kurz gefasst
Fachartikel in der Zeitschrift Canadian Biosystems Engineering, online erschienen am 31.12.2025. Verfasst von Joel Harms und 5 weiteren Autor:innen.
Abstract (Originalsprache)
Adapting grape varieties to specific environmental conditions—such as soil, climate, and topography—has long been a practice in viticulture. Over time, regions have been formally demarcated into appellations closely associated with particular grape varieties. However, with the accelerating impacts of climate change, wine-growing regions may soon face the decline of long-established varieties. To address this issue, the global diversity of grape varieties needs to be leveraged, but guidance is needed to select varieties appropriate for future climates. To address this challenge, Climate-Based Grape Variety Recommender Systems have been developed to help growers select the most suitable grape varieties for current and future climates. These systems rank the most widely known grape varieties for any given climate. In our research, a recommender system based on a coupled autoencoder (CAE) model was developed and benchmarked against similar models used for other crops. The model ranks 1,300 grape varieties by both popularity and suitability and was validated using global data and a case study in the wine regions of British Columbia. Results show that the CAE model significantly outperformed prior approaches. Moreover, the model’s suitability and popularity rankings align well with the trending and specialized grape varieties already favoured in different subregions of British Columbia. This approach offers a powerful, data-driven tool to support growers in making informed decisions under changing climate conditions—whether for established or emerging wine regions.
Zitieren
APA
Harms, J., Demers, L. E., Adamowski, J. F., Newlands, N. K., Adamchuk, V. I., & Castellarin, S. (2025). Guiding wine-grape variety selection using AI recommender systems. Canadian Biosystems Engineering, 67(1), 701. https://doi.org/10.7451/cbe.2025.67.7.1
Vancouver
Harms J, Demers LE, Adamowski JF, Newlands NK, Adamchuk VI, Castellarin S. Guiding wine-grape variety selection using AI recommender systems. Canadian Biosystems Engineering. 2025;67(1):701. doi:10.7451/cbe.2025.67.7.1
BibTeX
@article{Harms2025,
author = {Harms, Joel and Demers, Louis E. and Adamowski, Jan F. and Newlands, Nathaniel K. and Adamchuk, Viacheslav I. and Castellarin, Simone},
title = {Guiding wine-grape variety selection using AI recommender systems},
journal = {Canadian Biosystems Engineering},
year = {2025},
volume = {67},
number = {1},
pages = {701},
doi = {10.7451/cbe.2025.67.7.1}
}
EUVIMED