XploRe Help : pacfplot

Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Library:	times
See also:	pacf acf acfplot fft invfft

Quantlet:		pacfplot
Description:		plots the (sample) partial autocorrelation function of the time series x. The plot starts with the partial autocorrelation phi_1 at lag length 1.

Reference(s):

Mills, Terence C. (1999): The Econometrics of Financial Time Series. Second Edition. Cambridge, p.8-37

Usage:	`pacfplot(x {, maxlag})`
Input:
	`x`	T x 1 vector of observed time series
	`maxlag`	(optional) integer; maximum lag length that will be calculated, default value is 30

Note:

If the observations x_t follow an AR(p) process, then for lags greater than p the variance of the sample partial autocorrelation is approximately 1/T. So, confidence bands are given by +/- 2*T^(-0.5). They will also be shown in the plot.

Example:

library("times") ; loads the quantlets from times library dax = read("dax.dat") ; monthly DAX 1979:1-2000:10 return = tdiff(log(dax)) ; generates the monthly return pacfplot(return^2,12)

Result:

Display that shows the partial autocorrelation function of the
squared monthly DAX returns up to 12 lags. The total number of
observations is T = 261 and thus the confidence bands are
+/- 0.12.

Author: R. Chen, C. Hafner, W. Haerdle, R. Schulz, 20010419 license MD*Tech