|
改进主成分分析与多元回归融合的汉丰湖水质评估及预测 |
Evaluation and Prediction of Water Quality in Hanfeng Lake Based on Improved Principal Component Analysis and Multivariate Regression Model |
|
DOI: |
中文关键词: 改进主成分分析 多元回归 水质评估 汉丰湖 三峡水库 |
英文关键词: Improved principal component analysis Multivariate regression Water quality evaluation Hanfeng Lake Three Gorges Reservoir |
基金项目:国家自然科学基金青年科学基金资助项目(61605205);国家重大科研仪器研制基金资助项目(51727812) |
|
摘要点击次数: 830 |
全文下载次数: 23 |
中文摘要: |
基于2015—2017年汉丰湖水质监测数据,采用改进主成分分析和多元回归相融合的评价方法对水环境质量状况进行评价。先对水质主要影响因素采用改进主成分分析作降维处理并计算主成分得分值,再对选定的主成分作多元线性回归处理得到水质预测回归模型,并用于研究区水质的评估预测。结果表明:选出的4个主成分因子其累积方差贡献率达到873%,实现了数据结构的简化;同时,改进主成分回归预测值总体上更趋近于实测值,其预测结果的相对误差最大值<4%,而常规方法预测结果的相对误差最大值接近10%,体现出该方法所建模型具有较高的预测精度。 |
英文摘要: |
Based on the water quality monitoring data in Hanfeng Lake from 2015 to 2017, the water environmental quality was evaluated by improved principal component analysis and multivariate regression model. First, the major influence factors were performed to reduce the dimension by improved PCA, the scores of the principal components were calculated. Then, the water quality prediction model was established by multiple linear regressions analysis on the selected principal components. The model was applied in evaluating and predicting the water quality in the study area. The results indicated that the cumulative variance of 4 selected principal component factors reached 87.3%. This simplified the data structure. Meanwhile, the prediction value from improved principal component regression was generally close to the measured value, the maximum relative error was less than 4%. With conventional method, the relative error was close to 10%, showing that the model had high prediction accuracy. |
查看全文
查看/发表评论 下载PDF阅读器 |
关闭 |