Deck 16: Analyzing and Forecasting Time-Series Data

Question 1

In a recent meeting,a manager indicated that sales tend to be higher during October,November,and December and lower in the spring.In making this statement,she is indicating that sales for the company are cyclical.

Answer 1

False

Question 2

Model specification is the process of determining how well a forecasting model fits the past data.

Answer 2

False

Question 3

To compare one value measured at one point in time with other values measured at different points in time,index numbers must be used.

Answer 3

True

Question 4

The time-series component that implies a long-term upward or downward pattern is called the trend component.

Question 5

Stock analysts have recently stated in a meeting on Wall Street that over the past 50 years there have been periods of high market prices followed by periods of lower prices but over time prices have moved upwards.Given their statement,stock prices most likely exhibit only trend and cyclical components.

Question 6

A stockbroker at a large brokerage firm recently analyzed the combined annual profits for all firms in the airline industry.One time-series component that may have been present in these annual data was a seasonal component.

Question 7

Two common unweighted indexes are the Paasche Index and the Laspeyres Index.

Question 8

Some stocks are referred to as cyclical stock because they tend to be in favor for several years and then out of favor for several years.This is a correct use of the term cyclical.

Question 9

In order for a time series to exhibit a seasonal component,the data must be measured in periods as short or shorter than quarterly.

Question 10

In order to identify a cyclical component in time-series data,one year of weekly data should be sufficient.

Question 11

Harrison Hollow,an upscale eatery in Atlanta,tracks its sales on a daily basis.Recently,the manager stated that sales over the past three weeks have been very cyclical.Given the data she has,this statement is not a reasonable one to make.

Question 12

One of the basic tools for creating a trend-based forecasting model is regression analysis.

Question 13

The forecasting interval is the unit of time for which forecasts are made.

Question 14

You are given the following linear trend model: F_t = 345.60 - 200.5(t).The forecast for period 15 is approximately -2,662.

Question 15

While virtually all time series exhibit a random component,not all time series exhibit other components.

Question 16

If a manager is planning for an expansion of the factory,a forecast model with a long-term planning horizon would probably be used.•

Question 17

The Gilbert Company chief financial officer has been tracking annual sales for each of the company's three divisions for the past 10 years.At a recent meeting,he pointed to the annual data and indicated that it clearly showed the seasonality associated with its business.Given the data,this statement may have been very appropriate.

Question 18

If the historical data on which the model is being built consist of weekly data,the forecasting period would also be weekly.

Question 19

A time-series graph shows that annual sales data have grown gradually over the past 10 years.Given this,if a linear trend model is used to forecast future years' sales,the sign on the regression slope coefficient will be positive.

Question 20

An annual time series cannot exhibit a seasonal component.

Question 21

In a time series with quarterly sales data,assume that the seasonal index for the summer quarter has been found to be 0.87,this can be interpreted to mean that sales tend to be 87 percent higher in the summer quarter when compared to the other quarters.

Question 22

In using simple linear regression to find the linear trend in an annual time series from 1990 to 2005,the values 1990,1991,etc.are used as the values of the independent variable t when the regression is conducted.

Question 23

The Baker's Candy Company has been in business for three years.The quarterly sales data for the company are shown as follows:
Based on these data,the seasonal index for quarters 3,7,and 11 is approximately 1.61

Question 24

You are given the following linear trend model: F_t = 345.60 - 200.5(t).This model implies that in year 1,the dependent variable had a value of 145.1.

Question 25

One of the disadvantages of a regression-based linear trend forecasting model is that the forecast errors are computed for time periods that were used in developing the forecasting model and thus do not truly measure the forecasting ability of the model.

Question 26

In measuring forecast errors,the MAD and the square root of the MSE will provide similar (but not identical)values,in that both provide a measure of the "typical" amount of error in forecasts.

Question 27

Gibson,Inc.is a holding company that owns several businesses.One such business is a truck sales company.To help in managing this operation,managers at Gibson have collected sales data for the past 20 years showing the number of trucks sold each year.They have then developed the linear trend forecasting model shown as follows: Based on this information,it appears that the time series has a strong positive linear trend component.

Question 28

Renton Industries makes replacement parts for the automobile industry.As part of the company's capacity planning,it needs a long-range total demand forecast.The following information was generated based on 10 years of historical data on total number of parts sold each year. Based on this information we can conclude that the linear trend model explains a significant proportion of the variation in the number of parts sold,because the p-value is much smaller than any reasonable α that we might use.

Question 29

A plot of the time series with time on the horizontal axis is an effective means of assessing whether the series is linear or nonlinear.

Question 30

The Baker's Candy Company has been in business for three years.The quarterly sales data for the company are shown as follows:
As a first step in computing a seasonal index,the four-period moving average corresponding to the mid-point between periods 2 and 3 is 3,350.

Question 31

Gibson,Inc.is a holding company that owns several businesses.One such business is a truck sales company.To help in managing this operation,managers at Gibson have collected sales data for the past 20 years showing the number of trucks sold each year.They have then developed the linear trend forecasting model shown as follows: Based on this information,the fitted value for year 1 is about 99.

Question 32

Forecast bias measures the average amount of error per forecast,so a positive value means that forecasts tended to be too low.

Question 33

In comparing two or more forecasting models,the MAD value is useful in determining how successful the models were in fitting historical data.

Question 34

It is possible to use linear regression analysis to develop a forecasting model for nonlinear data if we can effectively transform the data.

Question 35

A seasonal index is a statistic that is computed from time-series data to indicate the effect of the seasonality in the time-series data.

Question 36

The reason for using split samples in developing a forecasting model is to eliminate the potential for bias in the resulting model.

Question 37

The Baker's Candy Company has been in business for three years.The quarterly sales data for the company are shown as follows:
Based on this information,the data reflect both a linear trend and seasonal components.

Question 38

Renton Industries makes replacement parts for the automobile industry.As part of the company's capacity planning,it needs a long-range total demand forecast.The following information was generated based on 10 years of historical data on total number of parts sold each year. Based on this information,the percent of variation in the number of parts sold that is explained by the linear trend model is approximately 90.9.

Question 39

Renton Industries makes replacement parts for the automobile industry.As part of the company's capacity planning,it needs a long-range total demand forecast.The following information was generated based on 10 years of historical data on total number of parts sold each year. Based on this information,we can conclude that sales on average have been growing by more than 48 thousand annually.

Question 40

From an annual time series of a company's sales the linear trend model F_t = 127 + 54(t)has been developed.This means that on average sales have been increasing by 127 per year.

Question 41

If the Durbin-Watson d statistic has a value close to 2,there is reason to believe that there is no autocorrelation between the forecast errors.

Question 42

If the observed value in a time series for period 3 is y_t = 128,and the seasonal index that applies to period 3 is 1.20,then the deseasonalized value for period 3 is 153.6

Question 43

In a single exponential smoothing model,finding the forecast value for each period requires having the actual and forecast values from the proceeding period.This is not possible for the first period,so for the first period one should use the actual value as the forecast value.

Question 44

If you suspect that a nonlinear trend exists in your data,one way to deal with it in a trend-based forecasting application is to transform the independent variable,for example by squaring the time measure or maybe taking the square-root of the time measure.

Question 45

To deseasonalize a time series,assuming a multiplicative model,the observed values are divided by the appropriate seasonal index.

Question 46

If the forecast errors are autocorrelated,this is a good indication that the model has been specified correctly.

Question 47

The Durbin-Watson test for autocorrelation can be reliably applied to any sample sizes.

Question 48

Recently,a manager for a major retailer computed the following seasonal indexes:
Note that the index for Summer Qtr is missing.However,it can be determined that the index for that period is approximately 1.03

Question 49

If a forecasting model produces forecast errors (residuals)that are negatively correlated,then we expect a negative residual to be followed by another negative residual to be followed by another negative residual and so forth.

Question 50

An advantage of exponential smoothing techniques over a regression-based trend model is that the exponential smoothing model allows us to weigh each observation equally,thereby giving a fairer method of developing a forecast.

Question 51

If the Durbin-Watson test leads you to reject the null hypothesis,then you are concluding that the forecast errors are positively autocorrelated.

Question 52

The purpose of deseasonalizing a time series is that a strong seasonal pattern may make it difficult to see a trend in the time series.

Question 53

Herb Criner,an analyst for the Folgerty Company,recently gave a report in which he stated that the annual sales forecast based on 20 years of annual sales data was done using a seasonally adjusted,trend-based forecasting technique.Given the information presented here,this statement has the potential to be credible.

Question 54

When using the multiplicative time-series model to determine the seasonal indexes,the first step is to isolate the seasonal and random components from the cyclical and trend components.

Question 55

Recently,a manager for a major retailer computed the following seasonal indexes:
The manager then developed the following least squares trend model based on the past five years of quarterly data: = 200 + 11.5t.Based on this,the seasonally adjusted forecast for quarter 25,which is the winter quarter,is 489.11

Question 56

The reason for testing for the presence of autocorrelation in a regression-based trend forecasting model is that one assumption of the regression analysis is that the residuals are not correlated.

Question 57

Large values of the Durbin-Watson d statistic indicate that positive autocorrelation among the forecast errors exists.

Question 58

A seasonally unadjusted forecast is one that is made from seasonal data without any adjustment for the seasonal component in the time series.

Question 59

If a time series involves monthly data there will be a total of 12 seasonal indexes.

Question 60

It is possible to conduct a statistical test for autocorrelation using the Durbin-Watson test and not be able to make a definitive conclusion about whether there is autocorrelation or not based on the data.

Question 61

In a double smoothing model,large values for the two smoothing constants will result in greater smoothing of the time series.

Question 62

The Morgan Company is interested in developing a forecast for next month's sales.It has collected sales data for the past 12 months.
After analyzing these data,if the company wishes to use double exponential smoothing with alpha = 0.20 and beta = 0.20,the starting values for the constant process and the trend process can be derived from a linear trend regression model by using the intercept and slope coefficient respectively.

Question 63

If a time series contains substantial irregular movement,the smoothing constant for a single exponential smoothing model that is close to 1.0 will result in forecasts that are not as smoothed out as those that would occur if a smaller smoothing constant was used.

Question 64

If a time-series plot indicates that the data do not appear to exhibit a trend,then a double exponential smoothing model would likely be the most appropriate to use rather than simple exponential smoothing model.

Question 65

Because simple exponential smoothing models require a starting point for the first period forecast that will be arbitrary,it is important to have as much data as possible to dampen out the effect of the starting point.

Question 66

The owners of Hal's Cookie Company have collected sales data for the past 8 months.These data are shown as follows:
Using a smoothing constant equal to 0.20 and starting forecast in period 1 of 100,the forecast value for period 9 is approximately 104.2.

Question 67

Prior to conducting double exponential smoothing a simple linear regression is conducted and the trend equation is = 42 + 38.3t,so the smoothed constant process value should be C₀ = 38.3 and the smoothed trend value should be T₀ = 42.

Question 68

In a double smoothing model,the second smoothing constant is introduced to account for the trend in the data if one exists.

Question 69

The Morgan Company is interested in developing a forecast for next month's sales.It has collected sales data for the past 12 months.
After analyzing these data,if the company wishes to use exponential smoothing,it should employ a single smoothing model since there is evidence of a linear trend in the data.

Question 70

In a recent meeting,the marketing manager for a large hardware company stated that he needed to have a forecast prepared for the next three months.The three-month period is called:

A)the forecasting horizon.
B)the forecasting period.
C)the planning time.
D)a business cycle.

Question 71

In establishing a single exponential smoothing forecasting model,a starting point for the forecast value for period 1 is required.One method for arriving at this starting point is to use the first data point as the forecast for that period.If we do that,then the first data point should be ignored when computing measures of forecast error.

Question 72

The process of selecting the forecasting technique to use in a particular application is called:

A)model diagnosis.
B)model specification.
C)validity checking.
D)model fitting.

Question 73

The owners of Hal's Cookie Company have collected sales data for the past 8 months.These data are shown as follows:
Using a starting forecast in period 1 of 100,the forecast bias over periods 2-8 is negative when a single exponential smoothing model is used with a smoothing constant of 0.20

Question 74

If a smoothing model is applied with a smoothing constant exceeding 0.50,the forecasting bias will tend to be positive in most cases.

Question 75

In a single exponential smoothing model,a large value for the smoothing constant will result in greater smoothing of the data than will a smoothing constant close to zero.

Question 76

The Morgan Company is interested in developing a forecast for next month's sales.It has collected sales data for the past 12 months.
Assuming that the company plans to use double exponential smoothing with starting values for the smoothed constant process value and smoothed trend value of 98.97 and 13.16 respectively,the MAD value for periods 2-12 is greater when alpha = 0.20 and beta = 0.20 than when alpha = 0.10 and beta = 0.10.

Question 77

Double exponential smoothing is used instead of single exponential smoothing when extra smooth forecasts are desired.

Question 78

If you suspect that your time-series data has a strong downward trend,you should set the beta smoothing constant at value fairly close to negative 1.0

Question 79

In a single exponential smoothing model,one smoothing constant is used to weigh the historical data,and the model is of primary value when the data do not exhibit trend or seasonal components.

Question 80

The owners of Hal's Cookie Company have collected sales data for the past 8 months.These data are shown as follows:
Using a smoothing constant of 0.4,the forecast value for period 3 is 112.