Daily Arxiv

This page organizes papers related to artificial intelligence published around the world.
This page is summarized using Google Gemini and is operated on a non-profit basis.
The copyright of the paper belongs to the author and the relevant institution. When sharing, simply cite the source.

Knowledge-guided machine learning for county-level corn yield prediction under drought

Created by
  • Haebom

Author

Xiaoyu Wang, Yijia Xu, Jingyi Huang, Zhengwei Yang, Yanbo Huang, Rajat Bindlish, Zhou Zhang

A study on corn yield prediction based on KGML-SM

Outline

This study explores corn yield prediction using remote sensing technology. To overcome the limitations of existing process-based and machine learning models, we developed the KGML-SM model, which considers soil moisture as an intermediate variable, utilizing the Knowledge-Based Machine Learning (KGML) framework. To prevent overprediction in drought conditions, we designed a drought-aware loss function and experimentally demonstrated that the KGML-SM model outperforms other machine learning models. We analyzed the relationships among drought, soil moisture, and corn yield, providing interpretability of prediction errors and suggesting future directions for model improvement.

Takeaways, Limitations

•
Takeaways:
◦
Emphasize the importance of soil moisture and improve prediction accuracy in drought situations through the development of the KGML-SM model.
◦
Enhance the understanding of corn yield prediction models by providing interpretability of the models.
◦
Demonstrated superior performance compared to other machine learning models.
•
Limitations:
◦
Lack of description of the technical details of the specific model (e.g., the specific machine learning algorithm used, details of the dataset).
◦
Further research is needed to determine generalizability to other crops or regions.
◦
Further analysis of the effectiveness of drought-aware loss functions is needed.
👍
Made with Slashpage