【摘要】 湖泊富营养化的预防与治理已经成为了世界性的难题。湖泊富营养化评价是湖泊治理的基础,可以为治理工作提供科学的依据。本文利用自组织特征映射网络模型对干旱区草型浅水湖泊——乌梁素海的水环境状况进行了聚类分析研究,并与层次聚类分析法进行了综合比较。结果表明,两种方法得出的结果相似度较高,均能反映实际情况。最后综合分析两种方法的计算结果,得出能够客观、清晰地反映乌梁素海水环境现状的分类结果。为乌梁素海富营养化的综合治理提供了合理、科学的依据。SOM网络操作过程简单易行,只需将相关数据提供给网络,网络就会通过自身训练、学习、竞争,最后输出相应的分类数目及对应的分类结果。研究表明,SOM网络是分析水环境状况的一种快捷、简便的分析工具。更多还原

【Abstract】 This paper is aimed to present our investigation findings with the water quality of Wuliangsuhai Lake in summer based on SOM model.Taking Wuliangsuhai Lake as a case study subject and using self-developed feature map model,we have made an evaluation of the eutrophic status,trying to take eutrophication as one of the most important factors in processing the lake paludification.As is known,lake eutrophication assessment is one of basic research topics in hoping to provide scientific evidence for lakes’ function revitalization.At the same time,we have also introduced hierarchical cluster analysis method into our research so as to compare what we have found with the SOM model.The results of our investigation have shown that the two methods are highly similar to each other in reflecting the real status quo of the lake.And,finally,joining the results gained with the use of the two methods and comparing the clearly identified results,we have generalized reasonable and scientific evidence for ways to revitalize the function of the lake.In conclusion,the work we have done helps to produce the best-classified results to meet the public wishes in classifying water quality and make it much simpler to use the SOM-ANN toolbox in MATLAB 7.0 software.更多还原