Using a computer package, find the degree of differencing needed to make the thermostat sales time series stationary. Then use the computer package to identify a Box-Jenkins model for forecast...
Using a computer package, find the degree of differencing needed to make the thermostat sales time series stationary. Then use the computer package to identify a Box-Jenkins model for forecasting thermostat sales. Perform appropriate diagnostic tests a. unit root testing is scope of this book, but an interested reader ght start by looking at a paper by Dickey, Bell, and 8 ller (1986) after mastering the notation in Chapter of this text. Taking differences when it is not cessary to do so is called overdifferencing. This an happen when it is difficult to tell whether an ri b. Recall that in Example 8.3 we used the Box-Jenkins model that is equivalent to Holt's ginal (or previously differenced) time series is trend-corrected exponential smoothing to forecast the thermostat sales. Discuss the pros and using the model you found in part (a) versus the tationary. Howeve overdifferenc r, some researchers believe that ing is not a serious problem when the kins equivalent Holt's trend corrected Box-Jen exponential smoothing model for forecasting the thermostat sales. analyst's goal is forecasting rather than testing an economic theory 0.13 Consider the thermostat sales data in Example 8.3 and Example 10.9.