改进决策树模型支持下的饮用水水源地健康风险评价

为了解决传统水环境健康风险评价模型中存在的忽略污染物浓度和时效性问题,基于决策树的分类思想方法,引入熵权法确定各指标的权重值大小,提出决策树综合评价模型,并应用于2009年FZ市饮用水水源地水环境健康风险评价。结果表明,2009年该市饮用水水源地水环境健康风险值为6.215×10-5,处于Ⅲ级、中风险状态,小于传统模型的计算结果(7.833×10-5)。认为虽评价结果为同一风险等级状态,但研究建立的综合模型可提高评价的准确度和精度。

An improved decision tree model for health risk assessment of drinking water source

The concentrations of pollutants and timeliness have been ignored in traditional water environmental health risk assessment models.To solve these problems,an improved decision tree comprehensive evaluation model was put forward based on the theory of decision tree classification,and the entropy method was used to determine the pollutants' weights.The model was applied to health risk assessment of the potable water source in FZ City in 2009.The results show that the water environmental health risk coefficient of the potable water source in FZ City was 6.215 × 10-5,being at level III with a medium-degree risk.It was lower than that calculated by the traditional model,i.e.,7.833 × 10-5,which has a same-level risk.The proposed model improves the accuracy and precision of evaluation.