Polynomial distributed lag glms

Usage

pdlglm(formula, data, family, na.action=na.omit, ...)

Arguments

formula As for glm except for special handling of pdl() terms
data see glm
family see glm
na.action handling for missing values
... see glm

Description

Fits a generalised linear model to time series data with special handling for polynomial distributed lag terms. When a pdl() term is included in the model formula it is expanded by the print and summary methods so that coefficients, standard errors and covariances for the lagged values of the predictor are reported. The print and summary methods give the same output as for glm; the plot method produces a plot of the fitted lag structure.

Value

An object of class pdlglm

Author(s)

Thomas Lumley

See Also

lm,glm

Examples

xx<-rnorm(1000)
zz<-rnorm(1000)
yy<--rnorm(1000)+sin(zz)+lagged(xx,1)
m<-pdlglm(yy~pdl(zz,4,2)+sin(zz)+pdl(xx,4,4))
m
summary(m)
par(mfrow=c(2,1))
plot(m)