$1.$ Create a period number

column and columns for seasonal indexes, to use for a seasonalized linear regression forecasting model.

$2.$ Use TREND with the data from the first $3$ weeks to generate a seasonalized, trend$-$including forecast for Production for weeks $1$ through $4.$

$3.$ Compute MSE and MAPE for the forecast compared to actual in weeks $1$ through $3.$

MSE

$4.$ Draw a line chart that shows the historical data and the forecast.

$4.$ put graph below here. In this problem, you are going to try a linear fit forecast as we covered in class and then a simple modification that we did not cover in class. In column C$,$ create a linear forecast using the period number as the predictor. Use all data points for the TREND. Then in column D$,$ starting in period $5,$ use a linear fit that uses the previous $4$ actuals for the $\backslash ($ y $\backslash )$ range and the previous $4$ period numbers for the $\backslash ($ x $\backslash )$ values. In the TREND function, you will not have absolute cell references for this. Then copy this down to period $20.$ Draw a graph that shows the actuals and the two forecasts. What do you think about the modified forecast method?