根据可解释的机器学习模型预测流域中重金属的生态风险:在数据稀缺的框架下
Hong Chen1,2, Meiling Kong2, Zhenghua Wu2
1School of Environment and Energy Engineering, Anhui Jianzhu University, Hefei, 230601, China.
Environmental monitoring and assessment
|January 28, 2026
概括
使用SMOGN的合成数据生成显著改善了机器学习模型的准确性,用于预测流域生态系统中的重金属毒性. 这种方法加强了生态风险评估,并支持了污染控制战略.
科学领域:
- 环境科学 环境科学
- 生态毒理学 生态毒理学
- 数据科学数据科学数据科学
背景情况:
- 重金属污染威胁流域生态系统,需要准确的生态风险预测.
- 由于成本高昂,生态毒理学数据有限,阻碍了机器学习应用.
研究的目的:
- 通过使用合成数据增加毒性数据集来解决生态毒理学数据稀缺问题.
- 开发和比较用于预测重金属毒性 (Cr,Mn,Cu) 的机器学习模型.
- 解释主要的毒性驱动因素,并评估Chaohu湖盆地的生态风险.
主要方法:
- 使用高斯噪声回归的合成少数群体过量采样技术 (SMOGN) 来生成合成生态毒性数据.
- 开发并比较了随机森林 (RF),极端梯度提升 (XGBoost) 和支向量机 (SVM) 的回归模型.
- 使用Shapley添加式解释 (SHAP) 进行模型解释性和风险系数 (RQ) 进行生态风险评估.
主要成果:
- 合成数据增强大大提高了模型性能,RF R2从0.696增加到0.977.
- 所有测试的模型 (RF,XGBoost,SVM) 与原始数据集相比,增强数据的准确性更高.
- 生态风险评估显示, (Cr) 和铜 (Cu) 在湖盆地存在高风险.
结论:
- SMOGN是克服生态毒理学数据短缺的可行和有效方法.
- 增强数据集提高了机器学习模型的准确性,用于预测重金属毒性.
- 该研究提供了关键的数据和证据,用于管理Chaohu湖盆地的重金属污染.
相关概念视频
Ecological Disturbance
21.0K
An ecological disturbance is a temporary disruption in the environment resulting from abiotic, biotic, or anthropogenic factors, causing a pronounced change in an ecosystem. The impact of an ecological disturbance, which can depend on its intensity, frequency, and spatial distribution, plays a significant role in shaping the species diversity within the ecosystem.
21.0K
Bioequivalence Data: Statistical Interpretation
210
Body:The statistical interpretation of bioequivalence data is a significant aspect of pharmaceutical research. Bioequivalence refers to the absence of any significant difference in the rate and extent to which the active ingredient in pharmaceutical products becomes available at the site of drug action when administered at the same molar dose under similar conditions. This helps determine if different drug products have similar absorption rates, ensuring their interchangeability.Statistical...
210
Ecological Succession
21.5K
Ecological succession is influenced by the processes of facilitation, inhibition, and toleration. Facilitation occurs when early successional species create more favorable ecological conditions for subsequent species, such as enhanced nutrient, water, or light availability. In contrast, inhibition happens when early successional species create unfavorable ecological conditions for potential successive species, such as limiting resource availability. In some cases, later successional species...
21.5K
Ecological Niches
26.4K
All organisms have a position within an ecosystem. The complete set of living and nonliving factors—including food resources, climate, and terrain—that define the position of a given organism are collectively referred to as the organism’s ecological niche.
26.4K
Bonding in Metals
52.4K
Metallic bonds are formed between two metal atoms. A simplified model to describe metallic bonding has been developed by Paul Drüde called the “Electron Sea Model”.
52.4K
Simplified Synchronous Machine Model
772
The Synchronous Machine Model is a fundamental tool in analyzing and ensuring the transient stability of power systems. This model simplifies the representation of a synchronous machine under balanced three-phase positive-sequence conditions, assuming constant excitation and ignoring losses and saturation. The model is pivotal for understanding the behavior of synchronous generators connected to a power grid, particularly during transient events.
In this model, each generator is connected to a...
In this model, each generator is connected to a...
772


