Question

Create the following forecasts:

a 4-period moving average forecast,

an exponential smoothing forecast with alpha=0.2,

and a linear trend forecast for your data.

Calculate the MSE and MAPE for each.

Plot the original series, and each of the forecasts.

Create forecasts for 2017 Q1, 2017 Q2, 2017 Q3 and 2017 Q4 using each of these three techniques. If you had to choose one of these, which would you choose and why?

Given your data, and this time period, would you prefer to try some other type of forecast, or modify the data in some way? (You must explain your answer for full credit.)

Area	kansas
2005:Q1	88446
2005:Q2	89264
2005:Q3	90515
2005:Q4	93050
2006:Q1	93413
2006:Q2	95259
2006:Q3	95481
2006:Q4	95203
2007:Q1	94289
2007:Q2	95951
2007:Q3	96938
2007:Q4	102358
2008:Q1	98133
2008:Q2	100673
2008:Q3	101098
2008:Q4	99447
2009:Q1	99848
2009:Q2	96214
2009:Q3	97177
2009:Q4	98842
2010:Q1	98981
2010:Q2	102666
2010:Q3	104749
2010:Q4	105408
2011:Q1	105323
2011:Q2	107018
2011:Q3	108065
2011:Q4	108743
2012:Q1	109229
2012:Q2	109741
2012:Q3	109308
2012:Q4	108626
2013:Q1	113688
2013:Q2	113875
2013:Q3	115014
2013:Q4	115249
2014:Q1	116195
2014:Q2	117594
2014:Q3	118550
2014:Q4	119348
2015:Q1	117064
2015:Q2	118592
2015:Q3	119560
2015:Q4	119492
2016:Q1	119920
2016:Q2	120704
2016:Q3	120683
2016:Q4	121448

Answer 1

I will choose the forecasting using linear trend as this method has the lowest MAPE and MSE