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: armacls armalik

Quantlet: arimaf
Description: prediction of the ARIMA(p,d,q) processes with known coefficients and d = 0,1

Usage: yp = arimaf(y,a,b,p,d,q,L{,constant})
Input:
y input series with at least p observations
a q x 1 vector, innovation series with q observations
b vector of the parameters of the ARIMA process. If the model has a constant, it must be represented as first element of the vector, the next p elements are the coefficients of the AR part and the last q elements are the coefficients of the MA part.
p scalar, number of AR lags of the ARIMA process
q scalar, number of MA lags of the ARIMA process
d scalar, number of differences, d=0,1
L scalar, forecast horizon
constant optional string, a constant is used if constant = "constant"
Output:
yp L x 1 vector of point forecasts

Note:

Example:
library("times")                        ; loads the quantlets from times library
L = 5                                   ; forecast horizon
randomize(0)                            ; sets a random seed
a = normal(145)                         ; innovation
b = #(0.9,0.7,-0.5)                     ; parameters of the process
mu = b[1]/(1-b[2])                      ; mean
y = genarma(b[2],b[3],a) + mu           ; ARMA(1,1) series
yf = arimaf(y,a[145],b,1,0,1,L,"constant")
yf

Result:
; Prediction of the process for the next five periods:

Contents of yf
[1,]   2.4824
[2,]   2.6377
[3,]   2.7464
[4,]   2.8225
[5,]   2.8757



Author: G. Moral, 20010608
(C) MD*TECH Method and Data Technologies, 05.02.2006