19Which method could provide more satisfactory long term forecast? AAMultiple regression modelBSimple moving averageCWeighted moving averageDSingle exponential smoothing20 Which of the following statements about autocorrelation of a time series with 50observations are correct? DaThe sampling distribution of autocorrelation is normal with mean zeroand variance 1/50.57

bThe lag-2 autocorrelation has a smaller absolute value than the lag-1autocorrelation.cThe Ljung-Box test can be used to test whether the lag-1autocorrelation is zero. dWhen the true autocorrelation is zero, the sample autocorrelationfunction can be negative.A(i), (iv)B(i), (ii), (iii)C(ii), (iii), (iv)D(iii), (iv)21 The quarterly sales for the Whole Foods Market, Inc. ($M) between 2004 and2007 are given in the following table. Which set of answers gives the correctseasonal indices for the four seasons? B2004200520062007Q11118.1481368.3281666.9531870.731Q2902.1411085.1581311.521463.21Q3917.3551132.7361337.8861514.42Q4927.3061115.0671291.0171743.412A0.74, 0.91, 1.08, 1.27B1.16, 0.92, 0.94, 0.98C1.02, 0.98, 1.08, 0.92D0.75, 1.25, 0.91, 1.09Answer:2004200520062007TotalSum of first four numbers *4/5191.347Q11118.1481368.3281666.9531870.7316024.161.160423297Q2902.1411085.1581311.521463.214762.0290.917301232Q3917.3551132.7361337.8861514.424902.3970.944340072Q4927.3061115.0671291.0171743.4125076.8020.977935399Sum=5191.347422 What time series model is most likely to generate data with the following PACFsand ACFs? C58

AMA(4)BMA(2)CAR(2)DAR(4)23Which statement concerning the forecast accuracy measure is true? BABias is usually greater than Mean Squared Error (MSE).BMean Absolute Error (MAE) depends on the measurement scale and unit ofthe time series.CMSE receives less influence from outliers than MAE.DMean Absolute Percentage Error (MAPE) depends on the measurement scaleand unit of the time series.24 The Holt’s exponential smoothing was applied to the American Express dailyclosing price time series (2ndcolumn) from 20031023 to 20031027 (year, month,day), where 20031025 is Saturday. The two smoothing parameters arealpha=0.2 and beta=0.8. Which of the following is incorrect according to thegiven information? DDateCloseLevel_t(L_t)Trend_t (D_t)Forecast2003102347.6347.85-0.2047.642003102447.6247.64-0.2147.432003102746.7547.47-0.1847.302003102847.542003102947.35AThe level value at 20031028 will be 47.19.BThe trend value at 20031028 will be -0.26.CThe forecast value at 20031029 will be 46.88.DThe forecast value at 20031030 will be 46.98.Answer:59

DateCloseLevel_t(L_t)Trend_t (D_t)Forecast2003101548.3048.300.002003101648.3848.300.0048.302003101747.8448.320.0148.332003102047.9248.23-0.0748.172003102148.4048.12-0.1048.012003102247.0448.09-0.0448.052003102347.6347.85-0.2047.642003102447.6247.64-0.2147.432003102746.7547.47-0.1847.302003102847.5447.19-0.2646.922003102947.3547.05-0.1646.882003103046.5246.98-0.0946.8925Which model generates a unit root process? DAARMA(1,1).BAR(1).CMA(1).DSingle exponential smoothing with alpha=1.26For an ARMA(p,q) process with p>0 and q>0, how should we determine the orderp? CACheck the ACFs.BCheck the PACFs.CCheck the AICs.DWhen q is given, choose p=q.27 The gamma parameter in Winter’s exponential smoothing is used to adjust theseasonal component. Which of the following is true? CAIf gamma is close to one, there is almost no seasonality effect for the data.

#### You've reached the end of your free preview.

Want to read all 78 pages?

- Spring '14
- Regression Analysis, Time series analysis, R Square