3) Quarterly demand for smartphones at a retailer is as shown. Forecast quarterly demand for year 5 using simple exponential smoothing with a=0.05, regression with multiplicative seasonal index, and Winter's method with a=0.05, B=0.1, and v=0.15. For Winter's method, use the parameters of the regression with multiplicative seasonal index values as initial estimates of the parameters. Evaluate MAD, MAPE, MSE, bias and TS to comment on the choice of forecasting method. Year Quarter Period Sales O 1 - 513 2 932 III 3 1,509 IV 1,902 2 - 5 693 6 1, 163 7 1,857 IV 8 2,469 3 - 9 346 10 1,439 III 11 2,271 IV 12 3,079 4 - 13 1,070 14 1,751 15 2,785 IV 16 3,613