Hi Rob,

Thanks for the won­der­ful fore­cast pack­age. I have a ques­tion about it. I would like to use the auto.arima method in the fore­cast pack­age to gen­er­ate a model. I would like to store the coef­fi­cients (e.g. ma1, sma1 etc.) gen­er­ated by the algo­rithm and use that in a java pro­gram to pre­dict the value at any given time. What equa­tion can I use to do the same? Thanks for your help.

I am try­ing to do an a series of out-​​of-​​sample forecast,i.e. a rolling regres­sion of the pred­i­tions uti­liz­ing the pack­age Fore­cast.
The goal is to use a model, such as ets to make a pre­dic­tion on a series in a ts for­mat , the series fore­casted is a sub­set of S.ts, S.ts starts at 1(1st obser­va­tion) and runs through obser­va­tion #3000. The Start =st, the fore­cast width(length of subset)=wd. Each time a fore­cast is made, I want to export that fore­cast to Excel and move for­ward St+i, where i might run from 1 to 100; and make a new fore­cast, and so on.
As I am not much of a pro­gram­mer, I had thought to cre­ate a func­tion, and then a loop rou­tine. Unfor­tu­nately, I can’t get my func­tion to work. the parts of the func­tion work, s is sim­ply the daily close of the S&P500 ori­en­tated ETF SPY.
I first read in S.ts, give st and wd a value
and hoped that if I entered some­thing like :
source(“Desktop/etspred.R”)
s.pred<-etspred.R()
that I would see a pre­dic­tion printed out.….
I would appre­ci­ate any help, sug­ges­tions, alter­na­tives for accom­plish­ing the same goal,etc
Thanks

Ets.pred<- function(fore)
{
s<-window(s.ts,start=st, end=st+wd)
tsp<-BoxCox(s,lambda=l)
tsp.ets <- ets(tsp)
ftsp.ets <- forecast.ets(tsp.ets,h=6)
fftsp=(ftsp.ets$mean)
fsp <- InvBoxCox(fftsp, lambda=l)
print(fsp)
}

Hi Rob

I have just got into the area — are there any peo­ple you would rec­om­mend to fol­low on Twit­ter or any other blogs to keep abreast of the field pleaase?

Thanks

Jon

Hi Rob

I have any ques­tion about using fore­cast pack­age for R in java pro­gram. Iwant to store result of fore­cast in an array. This is my code, it can’t stop run­ning and I get­ting error mas­sage.
What can I do to store it? Thanks for your help.
================================================
pub­lic class Arima {

                pub­lic sta­tic void
main(String[] args) {

                                Rengine re = new Rengine(Rargs, false,
null);

                                System.out.println(“Rengine cre­ated,
wait­ing for R”);

                                if (!re.waitForR()) {

                                System.out.println(“Cannot load
R”);

                                return;

                                }

                                re.eval(“library(forecast);”);

                                re.eval(“data<-scan(‘D:/timeseries.txt’,skip=1);”);

                                re.eval(“datats<-data;”);

                                /​/​
I use auto.arima func­tion to fore­cast my data for 12 peri­ods. but the  
                                /​/​ default period  of fore­cast
result is ten period. How can I set to get it? 

                                re.eval(“arima<-auto.arima(datats);”);

                                re.eval(“fcast<-forecast(arima);”);

                                REXP fs =
re.eval(“summary(fcast);”);

                                /​/​
I want to get result of fore­cast and returned it at an array

                                dou­ble[] fore­cast = fs.asDoubleArray();

                                for(int i=0; i<forecast.length; i++)

                                System.out.println(forecast[i]);

                re.end();

                }

}