Smart Sensing Nitrogen Prescriptions via Conformal Profit Bounds for Precision Agriculture

A framework for risk-aware nitrogen prescriptions in precision agriculture, evaluated on a leakage-free group split of 10,000 samples from a Kaggle yield dataset. The selected tree-ensemble model achieved RMSE 1.531 and MAE 1.214, with conformal intervals reaching empirical coverage of 0.912 at a 0.90 target. Authored by Abuzar Khan and last updated on 2026-04-27.

Use Cases

• Optimizing nitrogen application rates based on yield-response modeling and profit bounds.
• Evaluating model uncertainty using conformal prediction intervals for agricultural prescriptions.
• Implementing a risk-aware decision framework that allows abstention when uncertainty is excessive.

Strengths

• Model performance is quantified with specific metrics: RMSE 1.531 and MAE 1.214.
• Conformal prediction intervals achieved empirical coverage of 0.912 at a 0.90 target and 0.963 at a 0.95 target.
• The risk-aware strategy resulted in a lower mean nitrogen application (112.07) compared to the expected-profit optimization (121.34).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The primary data file is a 9.4 MB DOCX document, which may not be a standard data format for direct analysis.

Provenance

Source

figshare

Collection Method

Evaluation used a leakage-free group split of 10,000 samples from a Kaggle yield dataset.

Freshness

Last updated 2026-04-27 05:16:28; freshness should be verified.