我国湖泊环境健康指数的统计分析

发布时间：2017-12-31 11:19

本文关键词：我国湖泊环境健康指数的统计分析　出处：《黑龙江大学》2015年硕士论文　论文类型：学位论文

【摘要】：湖泊水质的评价是湖泊水体环境评价的基础部分,对湖泊水体环境做出科学和客观的评价是社会发展的需要,同时也是人类身体健康的需要.对湖泊水质做出评价时,选择一个合理的湖泊水质评定方法是十分重要的.本论文基于累积Logistic回归模型,对湖泊水质等级进行评定.通过对我国江苏省某湖泊的实际监测数据的统计分析,对湖泊水质等级建立关于PH、总磷含量(TP)、ORP、总氮(TN)含量等7个影响因素的累积Logistic模型,并对模型进行修正.把江苏省某湖泊的实际监测样本按照一定的原则取出一部分作为训练样本,用剩余的部分数据作为检验数据,训练样本用来估计模型中参数,然后利用所构建的累积Logistic回归模型对训练样本和检验样本的水质等级进行预测来检验模型,根据所得结果可知,对于训练样本和检验样本预测等级正判率分别是91.67%和90.28%.利用我们建立的湖泊水体质量等级评定累积Logistic模型对其它湖泊实际监测样本的等级进行预测,结果显示在对我国吉林省某湖泊水质等级的预测中正判率为88.68%.在没有得到正确判断的样本中,所得的预测结果也是处于原等级经验值相邻等级,说明模型对湖泊水质等级的评定有较好的评估作用,具有一定的实际意义和应用价值.
[Abstract]:The evaluation of lake water quality is the basic part of lake water environment evaluation. It is the need of social development to make scientific and objective evaluation of lake water environment. At the same time, it is also the need of human health. It is very important to choose a reasonable method to evaluate lake water quality. This paper is based on cumulative Logistic regression model. Based on the statistical analysis of the actual monitoring data of a lake in Jiangsu Province, the author established the PHand total phosphorus content (TPN) ORP for the water quality grade of the lake. Cumulative Logistic model of seven influencing factors such as total nitrogen (TN) content. And the model is modified. The actual monitoring samples of a lake in Jiangsu Province are taken out as training samples according to certain principles, and the remaining part data are used as test data. The training sample is used to estimate the parameters of the model, and then the cumulative Logistic regression model is used to predict the water quality of the training sample and the test sample. For training samples and test samples, the positive rate of prediction grade is 91.67% and 90.28.The cumulative Logistic model established by us for evaluating the quality of lake water body is applied to other lakes. The grades of international monitoring samples were predicted. The results show that the correct judgment rate is 88.68 in the prediction of the water quality grade of a lake in Jilin Province. In the sample that has not been correctly judged, the predicted result is also in the adjacent grade of the experience value of the original grade. It shows that the model has a good evaluation function on the evaluation of lake water quality, and has certain practical significance and application value.
【学位授予单位】：黑龙江大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：X824

【参考文献】