CBE Forecasting 5

Question 1

The Box-Jenkins methodology for ARIMA modeling follows which sequence of steps?

Accepted Answer

Identify → Estimate → Diagnose → Forecast

Answer

Box-Jenkins proceeds by identifying the model order via ACF/PACF, estimating parameters, diagnosing residuals, and then forecasting if diagnostics pass.

Question 2

The Partial Autocorrelation Function (PACF) is primarily used in ARIMA modeling to identify:

Accepted Answer

The order of autoregressive (AR) terms

Answer

The PACF cuts off sharply after lag p for an AR(p) process, making it the standard diagnostic for selecting the autoregressive order.

Question 3

A forecast combining multiple models typically outperforms individual models because:

Accepted Answer

It leverages diverse information and reduces variance without proportionally increasing bias

Answer

Forecast combination diversifies across different model structures and information sets, reducing overall error variance through a bias-variance trade-off.

Question 4

Which statement about out-of-sample forecast evaluation is correct?

Accepted Answer

Out-of-sample evaluation uses data not seen during model estimation

Answer

Out-of-sample testing uses a reserved holdout period after the estimation window to provide unbiased evidence of how the model will perform on new data.

Question 5

When MAPE (Mean Absolute Percentage Error) is undefined or problematic, the most likely reason is:

Accepted Answer

Actual values include zeros or near-zeros

Answer

MAPE divides errors by actual values, so any actual value at or near zero causes division by zero, making the metric undefined or infinitely large.

(CBE) Certified Business Economist Practice Test

(CBE) Certified Business Economist Practice Test

CBE Forecasting 5