Predictive Efficacy of Classification and Regression Models for Tissue-Residue Toxicity of

Rui Wang's dataset on figshare, last updated in 2026, contains machine learning predictions for the chronic developmental toxicity of brominated flame retardants (BFRs) in fish. The dataset is 13.6 MB and includes results from random forest and XGBoost models, which achieved a test-set accuracy of 0.75 and an R² of 0.94, respectively. Feature analysis identified body weight and chemical structure as key determinants of toxicity, with risk assessments applied to species including threatened and endangered fish and regions like the Pearl River Basin.

Use Cases

• Prioritizing pollutants for ecological risk assessment based on predicted tissue-residue toxicity.
• Comparing model performance for classification and regression tasks in environmental toxicology.
• Analyzing the influence of fish body weight and chemical structure on toxic potency.
• Assessing regional ecological risks, such as in the Pearl River Basin, driven by specific contaminants like BDE-209.

Strengths

• Includes validated machine learning models with reported performance metrics (test-set accuracy 0.75, R² 0.94).
• Analysis covers both common and threatened and endangered (T&E) fish species.
• Feature analysis identifies concrete determinants like body weight and highly halogenated multiring aromatic scaffolds.
• Risk assessment provides specific geographic context, identifying the Pearl River Basin as a high-risk region.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect geographic or source bias inherent to the underlying research and platform.

Provenance

Source

Rui Wang via figshare

Collection Method

Integrated regression- and classification-based machine learning models leveraging cross-validation and application domain analysis.

Time Range

null

Freshness

Last updated 2026-05-04 04:47:45; freshness should be verified.

Geography

Includes analysis of the Pearl River Basin and seven representative freshwater basins worldwide.