Borrar filtros
Borrar filtros

Get residuals from real data and estimated arima model

13 visualizaciones (últimos 30 días)
Vardan Petrosyan
Vardan Petrosyan el 21 de Mayo de 2016
Respondida: ABDULAZIZ ALTUN el 20 de Abr. de 2020
Hi all, I need to estimate ARIMA model parameters and get residuals from the real data and the ARIMA model. Is the infer right function to do that? Currently I am doing this:
initial_arima_model = arima(4,0,2);
arima_model = estimate(initial_arima_model,LEARN_DATA);
[E,V] = infer(arima_model,LEARN_DATA);
plot(LEARN_DATA);
hold on;
plot(LEARN_DATA + E);
hold off;
But I am not sure that this is correct.

Respuestas (2)

Gautam
Gautam el 25 de Mayo de 2016
Editada: Gautam el 25 de Mayo de 2016
Hi Vardan,
You can use the 'resid' function to get residuals from the data and the model.Refer the following documentation for more information about this function:
Regards,
Gautam
  1 comentario
ABDULAZIZ ALTUN
ABDULAZIZ ALTUN el 20 de Abr. de 2020
It does not work!
Mdl = regARIMA(1,0,1);
resid(Equity_US,Mdl)
It does not give the residuals!

Iniciar sesión para comentar.


ABDULAZIZ ALTUN
ABDULAZIZ ALTUN el 20 de Abr. de 2020
I first fitted the data using
% Fit an ARMA(1,1) model
Mdl = regARIMA(1,0,1);
% EstMdl = estimate(Mdl,Equity_US);
EstMdl =estimate(Mdl,Equity_US);
Once you fit the data, you will need to read the summary of this
summarize(EstMdl)
Once you read the summary you will have to write the model yourself
Mdl = regARIMA('Intercept', 0.08245, 'AR', {0.69916}, 'MA',-0.82486, 'Variance',0.038536);
Now that you have written the model, you can infer the residuals
E = infer(Mdl,Equity_US)

Categorías

Más información sobre Autocorrelated and Heteroscedastic Disturbances en Help Center y File Exchange.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by