similar to: stats/arima.c memory allocation

Hi, the function makeARIMA(), designed to construct some state space representation of an ARIMA model, uses a C function called getQ0, which can be found at the end of arima.c in R source files (library stats). getQ0 takes two arguments, phi and theta, and returns the covariance matrix of the state prediction error at time zero. The reference for getQ0 (cited by help(arima)) is:

Hello everyone, I'm doing some benchmark comparing Arima [1] and SVR on time series data. I'm using an out-of-sample one-step-ahead prediction from Arima using the "fitted" method [2]. Do someone know how to have a two-steps-ahead forecast timeseries from Arima? Thanks, Matteo Bertini [1] http://robjhyndman.com/software/forecast [2] AirPassengers example on page 5

Hi *, Firstly, thank you so much for your time to read my email. I am currently interested in how to use R to predict time series from models fitted by ARIMA. The package I used is basic stats package, and the method I used is predict.Arima. What I know is that ARIMA parameters are estimated by Kalman Filter, but I have difficulty in understanding how exactly maximum likelihood (ML) estimator

Here is my problem: Autoregressive models are very interesting in forecasting consumptions (eg water, gas etc). Generally time series of this type have a long history with relatively simple patterns and can be useful to add external regressors for calendar events (holydays, vacations etc). arima() is a very powerful function but kalman filter is very slow (and I foun difficulties of estimation)

I'm testing some modifications in arima.c. I've noticed that a big internal array of double (rbar) is usually sparse and I'd like to add an option to store it as key-value mapping. Is there a library function or some other approach already used inside the R core for key-value mappings? Thanks, Matteo Bertini

Hi, i'm beginner in Bayesian methods, I'm reading the documentation about dlm package and kalman filters, I'm looking for a example of transformation of ARIMA in a state space equivalent to use the dlm package and calcualte the hyperparameters. Someone can help me about it?. If it's possible with a arima(1,0,1) example, or more complex model. While I have more examples best for me.

Hi, I am trying to forecast a model using predict.Arima I found arima model for a data set: x={x1,x2,x3,...,x(t)} arima_model = arima(x,order=c(1,0,1)) I am forecasting the next N lags using predict: arima_pred = predict(arima_model,n.ahead = N, se.fit=T) If I have one more point in my series, let's say x(t+1). I do not want to recalibrate themodel, I just want to forecast the next N-1

Hello, Does anyone know how the parameter estimates are calculated for xreg variables when called as part of an arima() command, or know of any literature that provides this info? In particular, I was wondering if there is a quick way to compare different combinations of "xreg" variables in the arima() fit in the same way that you would in multiple regression (using AIC & R^2

I'm searching a method to estimate the hyper-parameters in arima models. I'm reading about r-inla package, but in the examples section only talk about the AR part of the arima, but i need help about the MA part too. I'm beginner in Bayesian methods, I'm reading the documentation about dlm package and kalman filters, but the computacional cost of inla i think is better, but only

Hola! I got problems using an objects returned from arima (in KalmanSmooth(my.ts, ModArima$model), because my.ts showed up to have storage mode "integer" (is.integer(my.ts was TRUE). Should storage.mode() of a ts be allowed to be integer, should ts() someplace say storage.mode(ts.out) <- "double", or maybe inside arima() storage.mode(x) <- "double"

Hi all, Forecasting from an arima model is easy with predict. But I can't manage to backcast : invent data from the model before the begining of the sample. The theory is easy : take your parameters, reverse your data, forecast, and then reverse the forecast I've tried to adapt the predict function to do that (i'm not sure that the statistical procedure is fine (with the residuals),

Good morning everyone, I have attached an Excel file that contains a macro from which I call and use R's auto.arima function to generate forecasts. The program runs perfectly and it gets me the results; however, those results are pretty unusual. I also tried using the auto.arima function directly in the R console and still get weird results. The results are shown in columns AB, AC and AD

Hi all, First: I have a small line of code I'm applying to a variable which will be placed in a matrix table for latex output of accuracy measures: acc.aarima <- signif(accuracy(forecast(auto.arima(tix_ts, stepwise=FALSE), h=365)), digits=3). The time series referred to is univariate (daily counts from 12-10-2010 until 5-8-2010 (so not 2 full periods of data)), and I'm working on

Hi, I have been using this website ( http://www.stat.pitt.edu/stoffer/tsa2/Rissues.htm http://www.stat.pitt.edu/stoffer/tsa2/Rissues.htm ) to help me to fit ARIMA models to my data. At the moment I have two possible methods to use. Method 1 If I use arima(ts.data, order=c(1,2,0), xreg=1:length(ts.data)) then the wrong value for the intercept/mean is given (checked on SPSS and Minitab) and

Hi everyone I've got a problem regarding the arima() and the ar() function for autoregressive series. I would simply like to combine them. To better understand my question, I first show you how I'm using these two functions individually (see file in the attachement). 1) apply(TSX,2, function(x) ar(na.omit(x),method="mle")$order # this function finds the optimal

I'm using the arima function to estimate coefficients and also using predict.Arima to forecast. This works nicely and I can see that the results are the same as using SAS's proc arima. I can also take the coefficent estimates for a simple model like ARIMA(2,1,0) and manually compute the forecast. The results agree to 5 or 6 decimal places. I can do this for models with and without

Hello all, I use the arima to get a model, i.e. fit = arima(x,order=c(1,0,0)) and I know I can get the following from fit via est.coef = coef(fig) est.aic = fit$aic std.err = sqrt(diag(vcov(fit))) t.stat = est.coef/std.err How can I get the following stat from arima? Pr(>|t|) r2 adjust_r2 rmse Thanks, Rebecca

Hello all, I have been using R's time series capabilities to perform analysis for quite some time now and I am having some questions regarding its reliability. In several cases I have had substantial disagreement between R and other packages (such as gretl and the commercial EViews package). I have just encountered another problem and thought I'd post it to the list. In this case,

Dear members of the list, I''m a beginner in R and I''m having some trouble with: "Error in optim(init[mask], armafn, method = "BFGS", hessian = TRUE, control = optim.control, : non-finite finite-difference value [8]" when running "arima". I''ve seen that some people have come accross the same problem:

>If you ***look at the code*** for arima you will see that ``%+%'' is >defined >in terms of a call to ``.Call()'' which calls ``R_TSconv''. So >apparently >R_TSconv is a C or Fortran function or subroutine in a ``shared >object library'' >or dll upon which arima depends. Hence to do anything with it you'll >need to get >that shared