基于FCM-LVQ网络模型的疏勒河流域水安全评价

疏勒河流域水安全状况对祁连山脉的生态安全具有重要影响。针对疏勒河流域水安全问题,提出一种环境-生态-监管-治理模型,构建出祁连山脉内陆河流域水安全评价指标体系,通过模糊C-均值聚类分析(FCM)法结合专家打分法对指标数据进行处理,运用学习向量量化(LVQ)网络模型得到疏勒河流域水安全评价等级,并与单纯使用LVQ神经网络和BP神经网络的评价结果进行对比,以验证评价模型的实用性。结果表明:疏勒河流域水安全状况2013年表现为不安全,2014-2016年表现为基本安全,2017-2019年表现为安全,整体呈现为逐渐上升的趋势,这与流域内实际情况是相符的。另外,FCM-LVQ网络模型在运行速度及评价结果精度上明显更优于另外两种网络模型,可在流域水安全评价中推广使用。

Water security evaluation of Shule River Basin based on FCM-LVQ network model

The water security of Shule River Basin has an important impact on the ecological security of Qilian Mountains.Aiming at the water security problem of Shule River Basin,an environment-ecology-supervision-governance model was proposed to construct a water safety evaluation index system for the inland river basin in Qilian Mountains.The index data was processed using fuzzy C-means(FCM)algorithm combined with expert scoring method,and then the water security of the Shule River Basin was evaluated using learning vector quantization(LVQ)network model.The results were compared with those of the models using LVQ neural network only and BP neural network only to verify the applicability of the proposed model.The results show that the water security status of the Shule River Basin was evaluated as“unsafe”in 2013,“quasi-safe”from 2014 to 2016,and“safe”from 2017 to 2019,which presented a gradual upward trend.This is consistent with the actual situation in the basin.In addition,the FCM-LVQ network model is obviously superior to the other two network models in terms of running speed and accuracy of evaluation results,so it is recommended for watershed water security evaluation.