Evaluating climatic variability's impact on milk yield across climate zones: A machine learning-based comparative study of Switzerland and Thailand
Unprecedented climatic fluctuations have raised concerns about their impact on milk productivity across climate zones. However, differences in datasets and modeling methods hinder direct comparisons. This study systematically compares conventional and modern machine learning models for milk yield pr...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-12-01
|
| Series: | Smart Agricultural Technology |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772375525004691 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Unprecedented climatic fluctuations have raised concerns about their impact on milk productivity across climate zones. However, differences in datasets and modeling methods hinder direct comparisons. This study systematically compares conventional and modern machine learning models for milk yield prediction and evaluates the influence of climatic variability in Switzerland (moderate climate) and Thailand (tropical climate) using the same analytical framework.Five models—linear regression, ridge regression, gradient boosting regression, AdaBoost, and LightGBM—are tested with various feature sets, including Day-in-Milk, individual and combined meteorological variables, and three lagged milk yield values to assess their predictive value. Model performance is further analyzed on sub-datasets stratified by breed and parity range to explore climate effects in more homogeneous groups.The results demonstrate that simple models perform comparably to complex ones, with gradient boosting regression consistently outperforming other boosting methods. AdaBoost shows the weakest performance. Across all scenarios, previous milk yield is a stronger predictor than short-term meteorological variables, suggesting that recent production trends already reflect key weather effects. This pattern also holds within homogeneous sub-datasets. The findings offer insights into modeling climate influence on milk production across different climate zones, supporting the development of data-driven prediction models adaptable to diverse dairy farming systems. |
|---|---|
| ISSN: | 2772-3755 |