lagplot function

Lagged Regression Plot

Computes and plots the nonparametric regression function of a time series against its various lags.

lagplot(x, lag.max = 6, deg = 1, nn = 0.7, method = c("locfit", "gam", "both")[1])

Arguments

• x: time series
• lag.max: maximum lag
• deg: degree of local polynomial, needed only for the locfit method
• nn: fraction of nearest data contained in a window, needed only for the locfit method
• method: Two methods for nonparametric estimation: "locfit" is the default which uses the local polynomial approach via the locfit library to estimate the conditional mean function of $E(X_t|X_{t-k}=x)$ for $1\le k \le lag.max$; Another method is GAM, via the mgcv library.

Returns

Side effects: The nonparametric lagged regression functions are plotted lag by lag, with the raw data superimposed on the plots.

References

"Time Series Analysis, with Applications in R" by J.D. Cryer and K.S. Chan

Author(s)

Kung-Sik Chan

Examples

set.seed(2534567)
par(mfrow=c(3,2))
y=arima.sim(n=61,model=list(ar=c(1.6,-0.94),ma=-0.64))
# lagplot(y)