详细信息
基于模式识别神经网络的水资源配置评价模型研究 被引量:8
Research on the Evaluation Model of Water Resources Configuration Based on Pattern Recognition Neural Network
文献类型:期刊文献
中文题名:基于模式识别神经网络的水资源配置评价模型研究
英文题名:Research on the Evaluation Model of Water Resources Configuration Based on Pattern Recognition Neural Network
第一作者:李志新
机构:[1]贵州理工学院土木工程学院
第一机构:贵州理工学院土木工程学院
年份:2018
卷号:0
期号:11
起止页码:61-66
中文期刊名:中国农村水利水电
外文期刊名:China Rural Water and Hydropower
收录:CSTPCD;;北大核心:【北大核心2017】;
基金:贵州省科学技术基金计划项目(黔科合基础[2016]1062);国家自然科学基金项目(项目批准号51508121);贵州省科技合作计划项目(黔科合LH字2016[7096])
语种:中文
中文关键词:模式识别;神经网络;水资源;配置
外文关键词:pattern recognition;neural network;water resources;configuration
摘要:针对传统水资源配置合理性评价方法中存在的缺陷,为了解决其在评价过程中确定参数或指标权重时较大主观任意性的问题,基于模式识别神经网络和水资源配置评价指标及分级标准构建了模式识别神经网络水资源配置评价模型,该模型数据通过在分级标准阈值区间以随机内插方法产生,以分类误判百分率和交叉熵为模型性能评价指标,模型经过设计实现,对其进行了训练及测试实验,并结合实例应用该模型进行了全国各省级行政区水资源配置评价。实验结果表明:模式识别神经网络水资源配置评价模型精度性能较高、分类能力优良,其训练集、验证集及测试集交叉熵误差分别为2.81×10^(-7)、3.07×10^(-7)、1.31×10^(-6),且其分类误判百分率都为0;进一步的实例分析进一步表明,模型应用于水资源配置评价实践中的合理可行性,评价结果分析则揭示了在水资源配置中存在的突出问题,提出了改进配置合理性的建议。
In order to solve the problem in the traditional water resource allocation rationality evaluation method and solve the problem of subjective arbitrariness in determining the parameter or index weight in the evaluation process,this paper constructs a pattern recognition neural network water resource allocation evaluation model based on the pattern recognition neural network and water resource allocation evaluation index and classification standard.The model,the model data is generated by the random interpolation method in the classification standard threshold interval,and the classification error percentage and cross entropy as the model performance evaluation index.The model has been developed and tested by the design.The experimental results show that the model recognition neural network water resource allocation evaluation model has high accuracy and excellent classification ability.The cross entropy errors of the training set,the validation set and the test set are 2.81×10^(-7),3.07×10^(-7) and 1.31×10^(-6) respectively,and the percentage of their classification misjudgments is 0 percent.Further example analysis further shows that the model should be used.It is reasonable and feasible for the water resources allocation evaluation practice,and the analysis of the evaluation results reveals the outstanding problems in the water resources allocation,and puts forward some suggestions to improve the rationality of the allocation of water resources.
参考文献:
正在载入数据...