誤差に系列相関がある場合の回帰分析

概要

論文の詳細を見る
Usually, the error terms in linear regression model are assumed to be independently, identically and normally distributed with zero mean and constant variance. However, in case of the regression model for the estimation of air pollutant concentration based on meteorological factors, the error terms are correlated in time series. In such a case, the estimated errors are evaluated so small that the ordinary linear regression model may be inapplicable. The purpose of this paper is to propose asymptotic tests of hypothesis and parameter estimates of the linear regression model containing autocorrelated error terms, and to apply the regression model to the estimation of the concentration of sulfur dioxides based on meteorological elements in order to test the validity of the proposed regression model. In this paper, the ordinary regression analyses are presented in frequency domains in order to eliminate the autocorrelation of error terms.
明治大学の論文
1973-00-00