基于函数型非参数方法的气温数据预测分析

摘要

基于函数型非参数核回归估计的方法分析安徽省1955年至2010年月度平均气温数据,建立函数型非参数回归模型,并对2010年气温数据进行实证研究.同时,与经典的非参数回归模型的预测结果相比,本文方法的预测均方误差明显优于经典的非参数回归方法,体现出函数型非参数模型的优越性.%Temperature as a continuous time series, is a function of time.This paper mainly analyzes the average monthly temperature of the AnHui Province from 1955 to 2010 based on the kernel regression estimation method, establish functional nonparametric model and forecast the temperature in 2010.At the same time, compared with the finite-dimensional nonparametric model to demonstrate the advantages of the functional nonparameter model.