基于分形流形学习的支持向量机空气污染指数预测模型  

Air Pollution Index Prediction Model of SVM Based on Fractal Manifold Learning

作　　者：李萍 倪志伟[1,3] 朱旭辉[1,3] 伍章俊[1,3] LI Ping;NI Zhiwei;ZHU Xuhui;WU Zhangjun(School of Management,Hefei University of Technology,Hefei 230009;School of Information Engineering,Fuyang Normal University,Fuyang 236041;Key Laboratory of Process Optimization and Intelligent Decision-Making,Ministry of Education,Hefei 230009)

机构地区：[1]合肥工业大学管理学院,合肥230009 [2]阜阳师范学院信息工程学院,阜阳236041 [3]合肥工业大学过程优化与智能决策教育部重点实验室,合肥230009

出　　处：《系统科学与数学》2018年第11期1296-1306,共11页Journal of Systems Science and Mathematical Sciences

基　　金：国家自然科学基金(91546108;71490725;71521001);安徽省自然科学基金(1708085MG169);安徽省高校自然科学研究重点项目(KJ2018A0556);安徽省教育厅人文社会科学项目(JS2017AJRW0135)资助课题

摘　　要：针对目前北京、上海和广州地区较严重空气污染问题,建立了基于分形流形学习的支持向量机空气污染指数预测模型.首先采用分形理论计算出空气污染数据集分形维数;其次根据分形维数,采用流形学习将高维空气污染数据集通过非线性映射嵌入到低维空间中,对空气污染数据集进行降维;最后建立基于高斯核的支持向量机预测模型对三地区空气污染指数进行预测.北京、上海和广州三地空气污染指数预测结果表明,该模型较传统预测模型,预测性能更优,具有良好的稳定性和有效性.Air pollution index prediction model of SVM based on fractal manifold learning was proposed, for the serious air pollution of Beijing, Shanghai and Guangzhou areas at present. Firstly, the intrinsic dimension of air pollution data set is attained using fractal dimension;Secondly, the high dimension air pollution data set is embedded into a low-dimensional space using nonlinear mapping of manifold learning based on fractal dimension, which can reduce the dimension of the set;Finally, air pollution index prediction model of SVM based on Gaussian kernel function is built, which is applied in forecasting the air pollution index. Experimental results on the three data sets show that the prediction model is superior to other traditional models, and that it has high stability and effectiveness.

关 键 词：空气污染 流形学习 分形维数 支持向量机 

分 类 号：X51[环境科学与工程—环境工程] TP181[自动化与计算机技术—控制理论与控制工程]