Forecasting Chapter 15 15-1
Chapter Topics Forecasting Components Time Series Methods Forecast Accuracy Time Series Forecasting Using Excel Time Series Forecasting Using QM for Windows Regression Methods 15-2
Forecasting Components A variety of forecasting methods are available for use depending on the time frame of the forecast and the existence of patterns. Time Frames: Short-range (one to two months) Medium-range (two months to one or two years) Long-range (more than one or two years) Patterns: Trend Random variations Cycles Seasonal pattern 15-3
Forecasting Components Patterns (1 of 2) Trend - A long-term movement of the item being forecast. Random variations - movements that are not predictable and follow no pattern. Cycle - A movement, up or down, that repeats itself over a lengthy time span. Seasonal pattern - Oscillating movement in demand that occurs periodically in the short run and is repetitive. 15-4
Forecasting Components Patterns (2 of 2) Figure 15.1 (a) Trend; (b) Cycle; (c) Seasonal; (d) Trend w/season 15-5
Forecasting Components Forecasting Methods Times Series - Statistical techniques that use historical data to predict future behavior. Regression Methods - Regression (or causal ) methods that attempt to develop a mathematical relationship between the item being forecast and factors that cause it to behave the way it does. Qualitative Methods - Methods using judgment, expertise and opinion to make forecasts. 15-6
Forecasting Components Qualitative Methods Jury of executive opinion, a qualitative technique, is the most common type of forecast for long-term strategic planning. Performed by individuals or groups within an organization, sometimes assisted by consultants and other experts, whose judgments and opinions are considered valid for the forecasting issue. Usually includes specialty functions such as marketing, engineering, purchasing, etc. in which individuals have experience and knowledge of the forecasted item. Supporting techniques include the Delphi Method, market research, surveys, etc. 15-7
Time Series Methods Overview Statistical techniques that make use of historical data collected over a long period of time. Methods assume that what has occurred in the past will continue to occur in the future. Forecasts based on only one factor - time. 15-8
Time Series Methods Moving Average (1 of 5) Moving average uses values from the recent past to develop forecasts. This dampens or smoothes random increases and decreases. Useful for forecasting relatively stable items that do not display any trend or seasonal pattern. Formula for: n D = i = 1 i MA n n where: n = number of periods D = data in period i i in the moving average 15-9
Time Series Methods Moving Average (2 of 5) Example: Instant Paper Clip Supply Company forecast of orders for the month of November. Three-month moving average: 3 =i D i MA = 1 3 3 Five-month moving average: = 90 + 110 + 110 orders 3 130 = MA 5 5 =i D i = 1 5 = 90+ 110+ 130+ 75+ 50= 91orders 5 15-10
Time Series Methods Moving Average (3 of 5) Table 15.2 Three- and Five-Month Moving Averages 15-11
Time Series Methods Moving Average (4 of 5) Figure 15.2 Three- and Five-Month Moving Averages 15-12
Time Series Methods Moving Average (5 of 5) Longer-period moving averages react more slowly to changes in demand than do shorter-period moving averages. The appropriate number of periods to use often requires trial-anderror experimentation. Moving average does not react well to changes (trends, seasonal effects, etc.) but is easy to use and inexpensive. Good for short-term forecasting. 15-13
Time Series Methods Weighted Moving Average In a weighted moving average, weights are assigned to the most recent data. n WMA = W D n i = 1 i i where W = the weight for period i, between 0% and 100% i W = i 1.00 Example: Paper clip company weights 50% for October, 33% for September, 17% for August: 3 WMA = W D = (.50)(90) + (.33)(110) + (.17)(130) = 103.4 orders 3 i = 1 i i Determining precise weights and number of periods requires trialand-error experimentation. 15-14
Time Series Methods Exponential Smoothing (1 of 11) Exponential smoothing weights recent past data more strongly than more distant data. Two forms: simple exponential smoothing and adjusted exponential smoothing. Simple exponential smoothing: F t + 1 = αd t + (1 - α)f t where: F t + 1 = the forecast for the next period D t = actual demand in the present period F t = the previously determined forecast for the present period α = a weighting factor (smoothing constant). 15-15
Time Series Methods Exponential Smoothing (2 of 11) The most commonly used values of α are between 0.10 and 0.50. Determination of α is usually judgmental and subjective and often based on trial-and -error experimentation. 15-16
Time Series Methods Exponential Smoothing (3 of 11) Example: PM Computer Services (see Table 15.4). Exponential smoothing forecasts using smoothing constant of.30. Forecast for period 2 (February): F 2 = α D 1 + (1- α)f 1 = (.30)(.37) + (.70)(.37) = 37 units Forecast for period 3 (March): F 3 = α D 2 + (1- α)f 2 = (.30)(.40) + (.70)(37) = 37.9 units 15-17
Time Series Methods Exponential Smoothing (4 of 11) Table 15.4 Exponential Smoothing Forecasts, α =.30 and α =.50 15-18
Time Series Methods Exponential Smoothing (5 of 11) The forecast that uses the higher smoothing constant (.50) reacts more strongly to changes in demand than does the forecast with the lower constant (.30). Both forecasts lag behind actual demand. Both forecasts tend to be consistently lower than actual demand. Low smoothing constants are appropriate for stable data without trend; higher constants appropriate for data with trends. 15-19
Time Series Methods Exponential Smoothing (6 of 11) Figure 15.3 Exponential Smoothing Forecasts 15-20
Time Series Methods Exponential Smoothing (7 of 11) Adjusted exponential smoothing: exponential smoothing with a trend adjustment factor added. Formula: AF t + 1 = F t + 1 + T t+1 where: T = an exponentially smoothed trend factor T t + 1 + β(f t + 1 - F t ) + (1 - β)t t T t = the last period trend factor β = smoothing constant for trend ( a value between zero and one). Reflects the weight given to the most recent trend data. Determined subjectively. 15-21
Time Series Methods Exponential Smoothing (8 of 11) Example: PM Computer Services exponential smoothed forecasts with α =.50 and β =.30 (see Table 15.5). Adjusted forecast for period 3: T 3 = β(f 3 - F 2 ) + (1 - β)t 2 = (.30)(38.5-37.0) + (.70)(0) = 0.45 AF 3 = F 3 + T 3 = 38.5 + 0.45 = 38.95 15-22
Time Series Methods Exponential Smoothing (9 of 11) Table 15.5 Adjusted Exponentially Smoothed Forecast Values 15-23
Time Series Methods Exponential Smoothing (10 of 11) Adjusted forecast is consistently higher than the simple exponentially smoothed forecast. It is more reflective of the generally increasing trend of the data. 15-24
Time Series Methods Exponential Smoothing (11 of 11) Figure 15.4 Adjusted Exponentially Smoothed Forecast 15-25
Time Series Methods Linear Trend Line (1 of 5) When demand displays an obvious trend over time, a least squares regression line, or linear trend line, can be used to forecast. Formula: y = a+ bx where: a = intercept (at period 0) b= slope of the line x= the time period y = forecast for demand for period x b= xy nxy x2 nx a = y bx where: n= number x= n x y = n y of periods 15-26
Time Series Methods Linear Trend Line (2 of 5) Example: PM Computer Services (see Table 15.6) x= 12 78 = 6. 5 b= xy nxy 2 2 x nx y = 557 12 = 46. 42 = 3, 867 ( 12)( 6. 5)( 46. 42) = 1. 72 650 12( 6. 5) 2 a = y bx= 46. 42 ( 1. 72)( 6. 5) = 35. 2 y = 35. 2+ 1. 72x linear trend line for period 13, x = 13, y = 35. 2+ 1. 72( 13) = 57. 56 15-27
Time Series Methods Linear Trend Line (3 of 5) Table 15.6 Least Squares Calculations 15-28
Time Series Methods Linear Trend Line (4 of 5) A trend line does not adjust to a change in the trend as does the exponential smoothing method. This limits its use to shorter time frames in which trend will not change. 15-29
Time Series Methods Linear Trend (5 of 5) Figure 15.5 Linear Trend Line 15-30
Time Series Methods Seasonal Adjustments (1 of 4) A seasonal pattern is a repetitive up-and-down movement in demand. Seasonal patterns can occur on a quarterly, monthly, weekly, or daily basis. A seasonally adjusted forecast can be developed by multiplying the normal forecast by a seasonal factor. A seasonal factor can be determined by dividing the actual demand for each seasonal period by total annual demand: S i =D i / D 15-31
Time Series Methods Seasonal Adjustments (2 of 4) Seasonal factors lie between zero and one and represent the portion of total annual demand assigned to each season. Seasonal factors are multiplied by annual demand to provide adjusted forecasts for each period. 15-32
Time Series Methods Seasonal Adjustments (3 of 4) Example: Wishbone Farms Table 15.7 Demand for Turkeys at Wishbone Farms S 1 = D 1 / D = 42.0/148.7 = 0.28 S 2 = D 2 / D = 29.5/148.7 = 0.20 S 3 = D 3/ D = 21.9/148.7 = 0.15 S 4 = D 4 / D = 55.3/148.7 = 0.37 15-33
Time Series Methods Seasonal Adjustments (4 of 4) Multiply forecasted demand for entire year by seasonal factors to determine quarterly demand. Forecast for entire year (trend line for data in Table 15.7): y = 40.97 + 4.30x = 40.97 + 4.30(4) = 58.17 Seasonally adjusted forecasts: SF 1 = (S 1 )(F 5 ) = (.28)(58.17) = 16.28 SF 2 = (S 2 )(F 5 ) = (.20)(58.17) = 11.63 SF 3 = (S 3 )(F 5 ) = (.15)(58.17) = 8.73 SF 4 = (S 4 )(F 5 ) = (.37)(58.17) = 21.53 15-34
Forecast Accuracy Overview Forecasts will always deviate from actual values. Difference between forecasts and actual values referred to as forecast error. Would like forecast error to be as small as possible. If error is large, either technique being used is the wrong one, or parameters need adjusting. Measures of forecast errors: Mean Absolute deviation (MAD) Mean absolute percentage deviation (MAPD) Cumulative error (E bar) Average error, or bias (E) 15-35
Forecast Accuracy Mean Absolute Deviation (1 of 7) MAD is the average absolute difference between the forecast and actual demand. Most popular and simplest-to-use measures of forecast error. Formula: D F MAD= t t n where: t = the period number D t = demand in period t F t = the forecast for period t n = the total number of periods 15-36
Forecast Accuracy Mean Absolute Deviation (2 of 7) Example: PM Computer Services (see Table 15.8). Compare accuracies of different forecasts using MAD: D F MAD= t t n = 53.41 = 4.85 11 15-37
Forecast Accuracy Mean Absolute Deviation (3 of 7) Table 15.8 Computational Values for MAD and error 15-38
Forecast Accuracy Mean Absolute Deviation (4 of 7) The lower the value of MAD relative to the magnitude of the data, the more accurate the forecast. When viewed alone, MAD is difficult to assess. Must be considered in light of magnitude of the data. 15-39
Forecast Accuracy Mean Absolute Deviation (5 of 7) Can be used to compare accuracy of different forecasting techniques working on the same set of demand data (PM Computer Services): Exponential smoothing (α =.50): MAD = 4.04 Adjusted exponential smoothing (α =.50, β =.30): MAD = 3.81 Linear trend line: MAD = 2.29 Linear trend line has lowest MAD; increasing α from.30 to.50 improved smoothed forecast. 15-40
Forecast Accuracy Mean Absolute Deviation (6 of 7) A variation on MAD is the mean absolute percent deviation (MAPD). Measures absolute error as a percentage of demand rather than per period. Eliminates problem of interpreting the measure of accuracy relative to the magnitude of the demand and forecast values. Formula: D F MAPD = t t D t = 53.103 or 10.3% 520. 41 = 15-41
Forecast Accuracy Mean Absolute Deviation (7 of 7) MAPD for other three forecasts: Exponential smoothing (α =.50): MAPD = 8.5% Adjusted exponential smoothing (α =.50, β =.30): MAPD = 8.1% Linear trend: MAPD = 4.9% 15-42
Forecast Accuracy Cumulative Error (1 of 2) Cumulative error is the sum of the forecast errors (E = e t ). A relatively large positive value indicates forecast is biased low, a large negative value indicates forecast is biased high. If preponderance of errors are positive, forecast is consistently low; and vice versa. Cumulative error for trend line is always almost zero, and is therefore not a good measure for this method. Cumulative error for PM Computer Services can be read directly from Table 15.8. E = e t = 49.31 indicating forecasts are frequently below actual demand. 15-43
Forecast Accuracy Cumulative Error (2 of 2) Cumulative error for other forecasts: Exponential smoothing (α =.50): E = 33.21 Adjusted exponential smoothing (α =.50, β =.30): E = 21.14 Average error (bias) is the per period average of cumulative error. Average error for exponential smoothing forecast: E e = t n = 49.31 = 11 4.48 A large positive value of average error indicates a forecast is biased low; a large negative error indicates it is biased high. 15-44
Forecast Accuracy Example Forecasts by Different Measures Table 15.9 Comparison of Forecasts for PM Computer Services Results consistent for all forecasts: Larger value of alpha is preferable. Adjusted forecast is more accurate than exponential smoothing. Linear trend is more accurate than all the others. 15-45
Time Series Forecasting Using Excel (1 of 4) Exhibit 15.1 15-46
Time Series Forecasting Using Excel (2 of 4) Exhibit 15.2 15-47
Time Series Forecasting Using Excel (3 of 4) Exhibit 15.3 15-48
Time Series Forecasting Using Excel (4 of 4) Exhibit 15.4 15-49
Exponential Smoothing Forecast with Excel QM Exhibit 15.5 15-50
Time Series Forecasting Solution with QM for Windows (1 of 2) Exhibit 15.6 15-51
Time Series Forecasting Solution with QM for Windows (2 of 2) Exhibit 15.7 15-52
Regression Methods Overview Time series techniques relate a single variable being forecast to time. Regression is a forecasting technique that measures the relationship of one variable to one or more other variables. Simplest form of regression is linear regression. 15-53
Regression Methods Linear Regression Linear regression relates demand (dependent variable ) to an independent variable. y = a+ bx a= y bx b= xy nxy 2 2 x nx where: x= n x = mean of y = n y = mean of the x data the y data 15-54
Regression Methods Linear Regression Example (1 of 3) State University Athletic Department. Wins 4 6 6 8 6 7 5 7 Attendance 36,300 40,100 41,200 53,000 44,000 45,600 39,000 47,500 x (wins) 4 6 6 8 6 7 5 7 49 y (attendance, 1,000s) xy x 2 36.3 145.2 16 40.1 240.6 36 41.2 247.2 36 53.0 424.0 64 44.0 264.0 36 45.6 319.2 49 39.0 195.0 25 47.5 332.5 49 346.7 2,167.7 311 15-55
Regression Methods Linear Regression Example (2 of 3) x= 49 = 6. 125 8 y = 346. 43 34 8 9 =. b= xy nxy = ( 2167,. 70 ( 8)( 6. 125)( 43. 34) = 4. 06 2 2 ( 311) ( 8)( 6. 125) 2 x nx a = y bx= 43. 34 (. 406)( 6. 125) = 18. 46 Therefore, y = 18. 46+ 4. 06x Attendance forecast for x = 7 wins is y = 18. 46+ 4. 06( 7) = 46. 88 or 46, 880 15-56
Regression Methods Linear Regression Example (3 of 3) Figure 15.6 Linear Regression Line 15-57
Regression Methods Correlation (1 of 2) Correlation is a measure of the strength of the relationship between independent and dependent variables. Formula: r = Value lies between +1 and -1. n xy x y n x2 x2 n y2 ( y) 2 Value of zero indicates little or no relationship between variables. Values near 1.00 and -1.00 indicate strong linear relationship. 15-58
Regression Methods Correlation (2 of 2) Value for State University example: r = (8)(2, 167.7) (49)(346.7) (8)(311) (49)(49) (8)(15,224.7) (346.7)2 =.948 15-59
Regression Methods Coefficient of Determination The Coefficient of determination is the percentage of the variation in the dependent variable that results from the independent variable. Computed by squaring the correlation coefficient, r. For State University example: r =.948, r 2 =.899 This value indicates that 89.9% of the amount of variation in attendance can be attributed to the number of wins by the team, with the remaining 10.1% due to other, unexplained, factors. 15-60
Regression Analysis with Excel (1 of 6) Exhibit 15.8 15-61
Regression Analysis with Excel (2 of 6) Exhibit 15.9 15-62
Regression Analysis with Excel (3 of 6) Exhibit 15.10 15-63
Regression Analysis with Excel (4 of 6) Exhibit 15.11 15-64
Regression Analysis with Excel (5 of 6) Exhibit 15.12 15-65
Regression Analysis with Excel (6 of 6) Exhibit 15.13 15-66
Multiple Regression with Excel (1 of 4) Multiple regression relates demand to two or more independent variables. General form: y = β 0 + β 1 x 1 + β 2 x 2 +... + β k x k where β 0 = the intercept β 1... β k = parameters representing contributions of the independent variables x 1... x k = independent variables 15-67
Multiple Regression with Excel (2 of 4) State University example: Wins Promotion ($) Attendance 4 6 6 8 6 7 5 7 29,500 55,700 71,300 87,000 75,000 72,000 55,300 81,600 36,300 40,100 41,200 53,000 44,000 45.600 39,000 47,500 15-68
Multiple Regression with Excel (3 of 4) Exhibit 15.14 15-69
Multiple Regression with Excel (4 of 4) Exhibit 15.15 15-70
Example Problem Solution Computer Software Firm (1 of 4) Problem Statement: For data below, develop an exponential smoothing forecast using α =.40, and an adjusted exponential smoothing forecast using α =.40 and β =.20. Compare the accuracy of the forecasts using MAD and cumulative error. Period 1 2 3 4 5 6 7 8 Units 56 61 55 70 66 65 72 75 15-71
Example Problem Solution Computer Software Firm (2 of 4) Step 1: Compute the Exponential Smoothing Forecast. F t+1 = α D t + (1 - α)f t Step 2: Compute the Adjusted Exponential Smoothing Forecast AF t+1 = F t +1 + T t+1 T t+1 = β(f t +1 - F t ) + (1 - β)t t 15-72
Example Problem Solution Computer Software Firm (3 of 4) Period D t F t AF t D t - F t D t - AF t 1 2 3 4 5 6 7 8 9 56 61 55 70 66 65 72 75 56.00 58.00 56.80 62.08 63.65 64.18 67.31 70.39 56.00 58.40 56.88 63.20 64.86 65.26 68.80 72.19 5.00-3.00 13.20 3.92 1.35 7.81 7.68 35.97 5.00-3.40 13.12 2.80 0.14 6.73 6.20 30.60 15-73
Example Problem Solution Computer Software Firm (4 of 4) Step 3: Compute the MAD Values ( ) D F MAD F = t t t n = D AF MAD( AF ) = t t t n 41. 7 97 = 5.99 = 37. 5.34 7 39 = Step 4: Compute the Cumulative Error. E(F t ) = 35.97 E(AF t ) = 30.60 15-74
Example Problem Solution Building Products Store (1 of 5) For the following data: Develop a linear regression model Determine the strength of the linear relationship using correlation. Determine a forecast for lumber given 10 building permits in the next quarter. 15-75
Example Problem Solution Building Products Store (2 of 5) Quarter 1 2 3 4 5 6 7 8 9 10 Building Permits, x 8 12 7 9 15 6 5 8 10 12 Lumber Sales (1,000s of bd ft), y 12.6 16.3 9.3 11.5 18.1 7.6 6.2 14.2 15.0 17.8 15-76
Example Problem Solution Building Products Store (3 of 5) Step 1: Compute the Components of the Linear Regression Equation. x= 92 10 92 = y = 128. 12 86 10 6 =. b= xy nxy = ( 1290,. 3) ( 10)( 9. 2)( 12. 86) = 1. 25 2 2 ( 932) ( 10)( 9. 2) 2 x nx a= y bx= 12. 86 ( 1. 25)( 9. 2) = 1. 36 15-77
Example Problem Solution Building Products Store (4 of 5) Step 2: Develop the Linear regression equation. y = a + bx, y = 1.36 + 1.25x Step 3: Compute the Correlation Coefficient. r = n xy x y n x2 x2 n y2 ( y) 2 r = ( 10)( 1170,. 3) ( 92)( 128. 6) ( 10)( 932) ( 92)( 92) ( 10)( 1810,. 48) ( 128. 6) 2 =. 925 15-78
Example Problem Solution Building Products Store (5 of 5) Step 4: Calculate the forecast for x = 10 permits. Y = a + bx = 1.36 + 1.25(10) = 13.86 or 1,386 board ft 15-79
15-80