Model's forecast accuracy evaluation

[Aibek Ulan]

Hi everyone,

I want to evaluate forecasting accuracy of ARIMA (1,1,1) model. I use only one variable, so it is univariate. I have variable X with 1000 observations.
On first iteration I want to estimate this model by using only 900 observations from the sample. Then I want to obtain one step ahead forecast for 901st observation, calculate forecasting error by using 901st actual observation, store this error. On next iteration I want to estimate the model by using 901 actual observations, obtain forecast for 902nd observation, obtain forecasting error by using 902nd actual observation, store this error and so on until the last iteration.
As the result, in the end I will have 100 forecasting errors and will be able to calculate RMSE for this model. I will appreciate a lot if someone can help me with the code for this loop.

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ for asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

Is this research or a class assignment? We don't generally help with homework.

You need a loop from 900 to 999, estimating the model from 1 to the index, and then predicting one beyond the index. Then save the predicted value into another variable.