How do I estimate the residual $\varepsilon_{t}$ of a Seasonal ARIMA model $\hat{Y}_t=\hat{\phi}{Y}_{t-1}+\hat{\Phi}{Y}_{t-12}+\varepsilon_{t}$?
If the MSE is 0.114, what does it mean?


You can calculate $\varepsilon_{t}$ as follows. $$\varepsilon_{t} = {Y}_t - \hat{Y}_t $$

Mean squared Error can be calculated as

$$MSE = \frac{1} {n} \sum\limits_{t=1}^n ({Y}_t - \hat{Y}_t)^2$$


$$MSE = \frac{1} {n} \sum\limits_{t=1}^n (\varepsilon_{t} )^2$$

where $\hat{Y}_t$ is the predicted value from your model, ${Y}_t$ is the actual values.

An MSE of $0$ would indicate that the model fits your data perfectly which happens very rarely in practice. You have a value of 0.114 which is close to $0$. I would say your model fits your data very well. I would consult any basic statistics book if you want to learn more about MSE.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.