Exam 15: Forecasting

A 24-hour coffee/donut shop makes donuts every eight hours. The manager must forecast donut demand so that the bakers have the fresh ingredients they need. Listed below is the actual number of glazed donuts (in dozens) sold in each of the preceding 13 eight-hour shifts. Date Shift Demand(dozens) June 3 Day 59 Evening 47 Night 35 June 4 Day 64 Evening 43 Night 39 June 5 Day 62 Evening 46 Night 42 June 6 Day 64 Evening 50 Night 40 June 7 Day 69 Forecast the demand for glazed donuts for the three shifts of June 8 and the three shifts of June 9.

Smoothing methods are more appropriate for a stable time series than when significant trend and/or seasonal variation are present.

The exponential smoothing forecast for any period is a weighted average of all the previous actual values for the time series.

All quarterly time series contain seasonality.

For a multiplicative time series model, the sum of the seasonal indexes should equal the number of seasons.

To make period-to-period comparisons more meaningful and identify trend, the time series should be deseasonalized.

Quarterly billing for water usage is shown below. Year Quarter 1 2 3 4 Winter 64 66 68 73 Spring 103 103 104 120 Summer 152 160 162 176 Fall 73 72 78 88 a.Find the seasonal index for each quarter. b.Deseasonalize the data. c.Find the trend line. d.Assume there is no cyclical component and forecast the summer billing for year 5.

Which of the following is a qualitative forecasting method?

Explain what conditions make quantitative forecasting methods appropriate.

Time series methods base forecasts only on past values of the variables.

A qualitative forecasting method that obtains forecasts through "group consensus" is known as the

Forecast errors

If data for a time series analysis is collected on an annual basis only, which component may be ignored?

The multiplicative model

Causal models

Linear trend is calculated as T_t = 28.5 + .75t. The trend projection for period 15 is

In order to forecast the attendance at an annual tennis tournament, a model has been developed which uses attendance from the previous year and the amount spent for advertising this year. From the years shown in the table, forecast the attendance for years 2-5 and calculate the forecast error. Year Attendance Advertising Expenditure Forecast Error 1 8363 750 2 9426 1250 3 9318 3200 4 10206 4500 5 11018 5600 The multiple regression model is Attendance = 6738 + .23($) + .25 (Attlag)

Gradual shifting of a time series over a long period of time is called

The mean squared error is influenced much more by large forecast errors than by small errors.

Explain how to use seasonal index values to create a forecast.