基于ISOMAP与Adaptive-Lasso的日平均气温预测模型  

Prediction Model of Daily Average Temperature Based on ISOMAP and Adaptive-Lasso

作　　者：许欢 朱华亮[2] 温华洋[2] XU Huan;ZHU Hualiang;WEN Huayang(Basic Department,Hefei Preschool Education College,Hefei 230011,China;Anhui Meteorological Information Center,Hefei 230031,China)

机构地区：[1]合肥幼儿师范高等专科学校基础部,安徽合肥230013 [2]安徽省气象信息中心,安徽合肥230031

出　　处：《安庆师范大学学报（自然科学版）》2020年第3期38-42,共5页Journal of Anqing Normal University(Natural Science Edition)

基　　金：国家社科基金年度项目(16BTJ023);安徽省教育厅自然科学重点项目(KJ2018A0964)。

摘　　要：气候变化是全球公认的环境困境,研究气象数据的回归预测模型对预防自然灾害有重要意义。应用等距特征映射(ISOMAP)方法对气象因素进行降维,采用Adaptive-Lasso方法建立日平均气温与主成分之间的回归模型,实证分析了合肥2017.1.1—2018.8.31日平均气温的回归预测问题。结果表明,通过ISOMAP与Adaptive-Lasso方法建立的回归模型能更好地对日平均气温进行回归预测。Climate change is a globally recognized environmental dilemma.Studying the regression prediction model is important for preventing natural disasters.In this paper,ISOMAP method is used to reduce dimensions of meteorological factors.Then,Adaptive-Lasso method is used to establish the regression model between daily average temperature and principal component.The regression prediction of daily average temperature in Hefei from 2017.1.1 to 2018.8.31 is analyzed.The results show that the regression model established by ISOMAP and Adaptive-Lasso method can better predict the daily average temperature.

关 键 词：气象数据 ISOMAP Adaptive-Lasso 回归分析 

分 类 号：O212.8[理学—概率论与数理统计]