贝叶斯公式与模糊识别耦合方法在河流健康评价中的应用

Application of Bayesian Formula and Fuzzy Recognition Coupling Method in River Health Evaluation

为准确评价河流的健康程度,提出了基于贝叶斯公式与模糊识别耦合方法,分析了单个河流健康评价指标属某个等级的概率,用最大似然分类准则判定单个河流健康评价指标的评价等级,引入组合赋权法与相对隶属度综合确定各指标的权重,并将其应用于某市周边6条河流健康评价中。结果表明,除河流Ⅰ的健康评价结果为0.734之外,其他五条河流的评价结果在[0.586,0.628]之间。这表明仅有河流Ⅰ的评价等级为良,河流Ⅱ、Ⅲ、Ⅳ、Ⅴ的评价等级均为中等,河流Ⅵ的的综合评价结果为0.586,评价等级为差,评价结果与实际情况相符。可见基于贝叶斯公式与模糊识别耦合方法合理、可行,不仅提高了河流样本集的权重精确度,且更好地处理了不确定信息。

In order to accurately evaluate the river health level, Bayesian formula and fuzzy recognition coupling method has been proposed. The probability of single river health evaluation index belonging to a certain level is analyzed. The maximum likelihood classification criterion is used to judge the evaluation level of a single river health evaluation index. The weight of each index is comprehensively determined by combination weighting approach and relative membership de- gree. The method is applied to health evaluation of 6 rivers around a city. The results show that the evaluation level of river T is 0. 734 ; evaluation levels of other 5 rivers are in the range of 0. 586 and 0. 628. It illustrates that only river I rea ches good grade while rivers Ⅱ,Ⅲ,ⅣandⅤ are in a medium grade, and the evaluation level of river Ⅵ is 0. 586 with belonging to poor grade. The evaluation result is consistent with the actual situation. Therefore, the proposed method is feasible and reasonable, which not only improves the weight precision of rivers sample set but also deals with uncertain in- formation better.