基于GA-SVM的土壤重金属污染评价研究
在研究区域采集227个土壤样品,对其Cu、Zn、Pb、Cd和As含量进行测定,并分别应用单因子指数法、GA-SVM模型和内梅罗综合污染指数法计算、评价了各采样点的土壤环境质量等级。结果表明:GA-SVM模型的最佳惩罚参数C为21.939,RBF核函数的最优参数g为12.995,均方百分比误差MSPE为1.3958,该模型对训练集(150个样本)、测试集的平均分类精度达到97.33%;GA-SVM模型对77个测试样本的土壤环境质量等级评价结果与单因子指数法的评价结果一致,与内梅罗综合污染指数法评价结果的变化趋势一致。...
Saved in:
Published in | 江西农业学报 Vol. 29; no. 6; pp. 116 - 120 |
---|---|
Main Author | |
Format | Journal Article |
Language | Chinese |
Published |
广西大学轻工与食品工程学院,广西南宁530004%广西财经学院管理科学与工程学院,广西南宁,530003%广西大学轻工与食品工程学院,广西南宁530004
2017
广西壮族自治区环境监测中心站,广西南宁530028%澳大利亚国立大学克劳福德公共政策学院,澳大利亚堪培拉,541004%北京大学南宁附属实验学校,广西南宁,530029 广西财经学院管理科学与工程学院,广西南宁530003 |
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | 在研究区域采集227个土壤样品,对其Cu、Zn、Pb、Cd和As含量进行测定,并分别应用单因子指数法、GA-SVM模型和内梅罗综合污染指数法计算、评价了各采样点的土壤环境质量等级。结果表明:GA-SVM模型的最佳惩罚参数C为21.939,RBF核函数的最优参数g为12.995,均方百分比误差MSPE为1.3958,该模型对训练集(150个样本)、测试集的平均分类精度达到97.33%;GA-SVM模型对77个测试样本的土壤环境质量等级评价结果与单因子指数法的评价结果一致,与内梅罗综合污染指数法评价结果的变化趋势一致。 |
---|---|
Bibliography: | 36-1124/S YIN Juan1,2, LI Guo-xiang1 , WANG Xiao-fei2,3. , DENG Qu-cheng4, YANG Ni1 , ZHOU Zhong-hua5 ( 1. College of Management Science and Engineering, Guangxi University of Finance and Economics, Nanning 530003, Chi- na; 2. College of Light Industry and Food Engineering, Guangxi University, Nanning 530004, China; 3. Environmental Monito- ring Center of Guangxi Zhuang Autonomous Region, Nanning 530028, China; 4. Crawford School of Public Policy, Australian National University, Canberra 541004, Australia; 5. Nanning Experimental School Affiliated to Peking University, Nanning 530029, China) A total of 227 soil samples were collected in the study area, and the contents of heavy metal elements Cu, Zn, Pb, Cd and As in these soil samples were measured. By using single-factor index method, GA-SVM (Support Vector Machine) model and Nemero' s comprehensive pollution index method, the soil environmental quality level of each sample was calculated and evaluated respectively. The results indicated that: the optimal pun |
ISSN: | 1001-8581 |
DOI: | 10.19386/j.cnki.jxnyxb.2017.06.25 |