Question

1. By applying a log transformation to a seasonal data series, we eliminate the seasonality leaving us random series.
   1. TRUE
   2. False
2. Suppose we eliminate regression for monthly sales is fitted sales = 500 + 50*Dec – 80*June. Where Dec and Jun are monthly 0/1 indicators variables. The estimated coefficients of -80 represents.
   1. The estimated sales differential between June and any month other than December
   2. A June trend effects in that sales will drop -80 for June in the year and -160 for June In the year 2 and so on
   3. The estimated sales differential between June and December
   4. The predicted sales for June
3. In a process is random then the best summary and forecast fir for the data is a horizontal line
   1. True
   2. False

Answer 1

By applying a log transformation to a seasonal data series, we eliminate the seasonality leaving us random series.
False

Correct answer : Seasonality is removed by differencing.

Suppose we eliminate regression for monthly sales is fitted sales = 500 + 50*Dec – 80*June. Where Dec and Jun are monthly 0/1 indicators variables. The estimated coefficients of -80 represents.
correct answer : The estimated sales differential between June and any month other than December

In a process is random then the best summary and forecast fir for the data is a horizontal line
False