If a process is ARIMA(0,d,q), number of significant correlations in ACF plot tells the value of q.

A. True

B. False

How to estimate d in ARIMA(p,d,q) model?

A. Take random guess and keep trying until you find the optimal solution.

B. First try d=0 and note the error. Then try d =1 and note the error and then try d=2 and not the error. whichever d gives you lowest error in ARIMA model, use that d.

C. Use ADF test or KPSS test to determine if d makes the time series stationary or not. If not, increment d by 1.

D. Use ACF and PACF to estimate approximate d.

Augmented Dickey Fuller Test is used to prove randomness of the residuals of a forecasting method.

A. True

B. False

Augmented Dickey Fuller Test is used to prove randomness of the residuals of a forecasting method.

A. True

B. False

What is the naïve method forecast of following time series (1,7,2,7,2,1) for period 7?

A. 7

B. 1

C. 2

D. 3/2

If the difference between each consecutive term in a time series is constant, we call it Drift Model.

True

False

If the difference between each consecutive term in a time series is random, we call it random walk model.

True

False

If data exhibits quarterly seasonality, what is the seasonal naïve method forecast of following time series (4,1,3,2,5,1,2) for period 8?

A. 3

B. 1

C. 5

D. 2

E. 4

33. What command allows sub setting (cutting the time series into a smaller time series) of a time series in R ?

A. subset

B. cut

C. window

D. view

Which method of measure error is NOT appropriate when forecasting temperature time series which can have a real zero value?

A. RMSE

B. MAPE

C. MAE

D. MASE