Part -1:
Trend lines for yearly data
The following time series data represent the yearly amounts spent on advertising (in millions of dollars) by a large toy company:
24.7, 28.0, 28.1, 31.0, 36.8, 39.6, 45.0, 43.4, 41.7
This series of data begins in year 1994 (i.e., time period t = 1 corresponds to 1994). Using regression analysis, a linear trend line of the form Tt = 22.32 + 2.61 t was fit to the data. Using this information, generate a forecast for the total yearly amount of money that will be spent on advertising in 2008.
Part -2:
The quarterly number of permits granted for building houses in a large city is seasonal (there tend to be more permits granted for construction during spring and summer months than during winter months).
The following table shows the quarterly seasonal indexes for building permits:
|
Quarter 1
|
Quarter 2
|
Quarter 3
|
Quarter 4
|
|
0.61
|
1.63
|
1.32
|
0.44
|
Based on several years of permit data, beginning in Quarter 1 of 2000(i.e., t= 1 corresponds to Quarter 1 of 2000), an analyst reported that the trend line that describes the quarterly permit data is of the form Tt = 620 + 12t. Assuming that a multiplicative model can be used to describe the building permit data, generate a forecast for the number of building permits that you expect will be granted in Quarter 3 of 2005.