基于PCA-SVR模型中国工业固废产生量预测研究

浏览次数: 10
  • 分享到:

摘要:

依据国家统计局及中国统计年鉴数据,选取国内生产总值(GDP)、工业增加值、财政收入、固定资产投资、原煤产量、原油产量、发电量、年末总人口、我国工业企业单位数量等9个指标作为输入指标,构建了PCA-SVR(主成分分析-支持向量回归)中国工业固废产生量预测模型.并与支持向量回归(Support Vector Regression,SVR)、岭回归(Ridge Regression,RDG)、决策树(Decision Tree,DT)、提升树回归(Gradient Boosting Regression,GBR)多种单一模型的预测结果进行比对.实验结果表明,PCA-SVR组合模型的平均绝对百分误差(MAPE)为0.0630,均方根误差(RMSE)为2.6718,其预测误差最小.

In this paper,based on the information from China Statistical Yearbook of 1980-2015,data of gross domestic product(GDP),industrial added value,fiscal revenue,fixed asset investment,output of raw coal,crude oil production,electricity generation,population at the Year-end,the number of industrial enterprises in our country were selected as input features.The PCA-SVR(principal component analysis-support vector regression)prediction model for solid waste production in China is established.It was compared with the prediction results of a variety of single models including Support Vector Regression(SVR),Ridge Regression(RDG),Decision Tree(DT)and Gradient Boosting Regression(GBR).The experimental results showed that the mean absolute percentage error(MAPE)and root mean square error(RMSE)of PCA-SVR model are 0.0630 and 2.6718 respectively,and the prediction error is the smallest.

作者:

刘炳春 齐鑫

Liu Bingchun;Qi Xin(School of Management,Tianjin University of Technology,Tianjin 300384,China)

机构地区:

天津理工大学管理学院

出处:

《betway官方app 学报:自然科学版》 CAS 北大核心 2020年第1期69-74,共6页

基金:

天津市教委社会科学重大项目(2017JWZD16)。

关键词:

工业固废产生量 PCA-SVR 预测 政策引导

industrial solid waste generation PCA-SVR forecast policy guidance

分类号:

X825 [环境科学与工程—环境工程]


基于PCA-SVR模型中国工业固废产生量预测研究.pdf

Baidu
map