Related Experiment Video
Updated: Dec 27, 2025
15:30
A Telemetric, Gravimetric Platform for Real-Time Physiological Phenotyping of Plant–Environment Interactions
Published on: August 5, 2020
12.3K
Discriminative algorithm approach to forecast Cd threshold exceedance probability for rice grain based on soil
Jun Yang1, Chen Zhao1, Junxing Yang1
1Center for Environmental Remediation, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing, 100101, China; University of Chinese Academy of Sciences, Beijing, 100049, China.
Environmental Pollution (Barking, Essex : 1987)
|March 1, 2020
Summary
This study developed a probabilistic model to predict cadmium (Cd) in rice, improving accuracy by considering soil properties. The model quantifies uncertainty in Cd uptake, aiding pollution assessment in contaminated fields.
Area of Science:
- Environmental Science
- Agricultural Science
- Risk Assessment
Background:
- Cadmium (Cd) uptake in rice is highly variable due to soil properties and genetics.
- Existing models struggle to accurately predict Cd concentration in rice grains.
- Quantifying this uncertainty is crucial for food safety and risk management.
Purpose of the Study:
- To develop a probabilistic forecasting model for predicting Cd concentration in rice grains.
- To characterize the uncertainty in the relationship between soil Cd and rice grain Cd.
- To improve model performance by incorporating soil properties and using a discriminative algorithm.
Main Methods:
- Utilized a logistic regression (LR) model, a discriminative algorithm, due to non-normal and interdependent soil properties.
- Compared LR model performance against generative algorithms like naive Bayes and quadratic discriminant analysis.
- Incorporated soil physicochemical properties as parameters in a multivariate model.
Main Results:
- The LR-based model demonstrated superior performance over generative models.
- The multivariate LR model showed a significant improvement of 4.1% compared to the univariate model.
- Predicted probabilities from the LR model strongly correlated with the true exceedance rate (R² = 0.949).
Conclusions:
- The probabilistic forecasting model effectively quantifies uncertainty in Cd uptake by rice.
- The LR-based model provides a novel approach for assessing Cd pollution in rice from contaminated soils.
- Soil properties' influence on Cd exceedance probability varies with Cd concentration and threshold levels.
