河北大学学报(自然科学版) ›› 2016, Vol. 36 ›› Issue (3): 307-311.DOI: 10.3969/j.issn.1000-1565.2016.03.014

• • 上一篇    下一篇

基于遗传算法寻优的SVR雾霾预测模型

宗晓萍1,2,武子瀚1,刘言1   

  • 收稿日期:2015-05-13 出版日期:2016-05-25 发布日期:2016-05-25
  • 通讯作者: 武子瀚(1988—),男,河北张家口人,河北大学在读硕士研究生.E-mail:543308335@qq.com
  • 作者简介:宗晓萍(1964-),女,河北蔚县人,河北大学教授,主要从事模式识别、智能控制及混合动态系统、机器人视觉伺服控制方向研究. E-mail:769085906@qq.com
  • 基金资助:
    国家自然科学基金资助项目(11271106)

Optimization SVR fog prediction model based on genetic algorithm

ZONG Xiaoping1,2,WU Zihan1,LIU Yan1   

  1. Electronic Information Engineering College, Hebei University, Baoding 071002, China
    Hebei University-Rockwell Automation Laboratory, Baoding 071002, China
  • Received:2015-05-13 Online:2016-05-25 Published:2016-05-25

摘要: 针对雾霾天气愈发严重及难以预测的问题,提出一种以GA(遗传算法)优化支持向量回归机(SVR)参数的预测模型.首先利用因子分析对气象因子降维,然后再通过GA对SVR的参数寻优,并把最优参数带入SVR模型,对保定PM2.5浓度进行预测.对比参数模型的预测结果,为雾霾预测选出一种新的模型.

关键词: PM2.5预测, SVR, 因子分析, GA

Abstract: To perform smog forecast,this paper puts forward a model,which is based on support vector regression(SVR)method and GA method.The first eigenvector dimension is reduced using factor analysis,and the SVR parameters are optimized through the GA method,and put into the SVR model for the prediction of the PM2.5 value of the city of Baoding.Through comparison, an optimized prediction model is obtained.

Key words: PM2.5 forecast, SVR, factor analysis, GA method

中图分类号: