Use this window to view all the models and parameters required for working in IFS/Demand Plan. You can move this window to any other part of the MDI window, or even outside the window, by clicking the Name field (on the left). The detail view can be turned on or off by using the View command on the Toolbar menu. Read-only fields have a gray background, whereas read or write fields have a white background. Any area of this window's contents may be selected and copied to the clipboard for pasting in any other appropriate system.
N/A
| Menu Choice | Activity |
| Properties | Displays a dialog box from which you can select the fields that should be visible. |
| Copy Down | Copies the forecast model to all the parts in its group. Displays the Copy down dialog box. |
| Copy forecast parameters | Copies the forecast parameters from used (e.g. Forecast Model) to Std forecast parameters (e.g. Std forecast model) |
| Format Cells | Cell formatting in the table is limited to specifying the number of decimal places to be used and whether to use the 1000 separator. |
A number of the fields in the Detail View appear twice, with one field having the Std prefix added to its name. (Example: Forecast Model and Std Forecast Model.) If you change the model or parameter with the Std prefix, you make the change permanent. This means that this model or parameter is used when the server updates/creates the next main forecast scenario (normally at the start of a new forecast period). Changing the model or parameter without the Std prefix, however, affects only the current selected forecast scenario and has no consequences for future forecast calculations.
Note On combined flow and on group selections it is only possible to change the Std fields. This is because these levels does not exist in the data base so there will not be any automatic forecast calculations on these levels.
The following models and parameters are presented in the window:
Scenario: Here you can alternate between the different scenarios that exist in this period. The main (0) scenario is always present, the other scenarios are named scenario 1, scenario 2 and so on.
Forecast Model:
The system provides eleven different forecast models:
| Manual | The value entered in Manual Yearly is spread out evenly through the year. |
| Moving Average | Sets all the periods of the forecast equal to the average of historical demand. The average is based on the number of most recent periods entered in the Moving Average Periods. |
| EWMA Level | Exponential Weighted Moving Average. |
| EWMA Level and Trend | Exponential smoothing model with trend enhancement. |
| Naive | The naive forecast model |
| AEWMA | AEWMA is the Adaptive Exponential Weighted Moving Average. |
| Regression | The least squares fitting line. This is the line that minimizes the sum of squared errors. |
| Best fit | Finds the forecasted model that best explains the historical pattern of the above models. |
| Browns Level And Trend | Exponential smoothing model with trend enhancement, with a single smoothing constant. |
| Croston Intermittent | Croston's intermittent model used for slow moving parts. |
| Multiple Regression | Forecast model uses one or more explanation variables to compute the forecast. |
| Bayesian | This is a good all-round forecast model, when you don't have the skill or you don't have the time to tune some of the other models, then this model is a good choice. A good/robust default forecast model. |
| Life Cycle | This model is used to forecast the phase in and/or the phase out parts of a parts life cycle. If the part is used in between these phases it will be like a EWMA Level Model. |
A detailed description of the models can be found in Forecast Models.
Alpha: An exponential smoothing constant for Level that determines how much emphasis will be given to recent demand data. It is used in the EWMA Level and EWMA Level and Trend models. A high alpha value gives more weightage to recent demand data, and so providing a more responsive forecast model that becomes unstable in the presence of randomness. Low values spread the weight over older demand data and result in more stable forecasts. Values between 0.01 and 0.3 are recommended under normal circumstances. High values should not be used except under close supervision. Values for alpha must be in the range 0.0-1.0.
Beta: An exponential smoothing constant for Trend that determines how much emphasis will be given to recent trend factors. It is used in the forecast model EWMA Level and Trend. A high value for beta gives more weight to recent trend factors and produces a forecast model that responds faster to changes in trends but is unstable in the presence of randomness. Values must be in the range 0.0-1.0.
Rho: A trend reduction constant. The trend factor in the EWMA Level and Trend forecast model is multiplied by this constant for each period in the forecast horizon. With Rho set to 1, the trend continues indefinitely into the forecast horizon. Values between 0.75 and 1.00 are recommended. Values must be in the range 0.0-1.0
Moving Average Periods: The number of periods of historical demand used to find the arithmetic average in the Moving Average forecast model. This field is available only when this forecast model is selected. A detailed description of the Moving Average forecast model can be found in Forecast Models.
Season Profile: Combo box for selecting the user-created seasonal profile for demand that is to be used for the current Part No, or aggregated group. Season Profiles may be used in combination with any of the forecast models, except the Naive model. A detailed description of profiles can be found in Season Indexes. Note that the automatic season profile will use the system season profile for the part, if the part has passed the seasonal component test. See Season Indexes help file for further explanation.
Note: The Std Season Profile field decides the profile that is displayed in the Season Profile Graph and the Season Profile Table.
Manual Yearly: The manually-entered yearly demand used in the Manual forecast model. Manual Yearly is always expressed in standard units, regardless of the selection made in Select Unit. This field is available only when the Manual forecast model is selected. A detailed description of the Manual Yearly forecast model can be found in Forecast Models.
Period Profile: Como box for selecting the user-defined period profile that is to be used of this part or group. A period profile defines how the daily forecast is distributed over the length of the period, it might be that the first 5 days of the period stands for 25% of the total sales within the period. Period profiles may be used with any of the forecast models. A detailed description can be found in Period Profile. This field is disenabled when you have weekly period version as the Demand Plan Servers main period version. When other period versions are used you can combine a profile for the entire forecast period and a weekly profile to form the resulting daily forecast.
Note: The Period Profile field decides the profile that is displayed in the Profile Graph and the Profile Table.
Weekly Profile: Como box for selecting the user-defined weekly profile that is to be used of this part or group. A week profile defines how the sales is distributed over a calendar week, it might be that Mondays only sells 8% of the total sales of the week, while a Saturday sells 30% of the total weekly sales. Weekly profiles may be used with any of the forecast models. A detailed description can be found in Weekly Profile.
Note: The Period Profile field decides the profile that is displayed in the Profile Graph and the Profile Table.
Delta: An exponential smoothing constant for the Tracking Signal. A detailed description of the tracking signal can be found in Forecast Measurements.
Lead Time: The current part's lead time.
Inheritance: Used to turn on and off inheritance of historical demand data from the inventory item entered in the Predecessor field in IFS/Demand Planner/Forecast Parts. See Forecast Part Info.
Predecessor: The Part No from which the current part is to inherit historical demand. This field can only be defined in IFS/Demand Planner/Forecast Parts see Forecast Part Info.
Inheritance Start: The start date from which historical data is to be inherited. This field can only be defined in IFS/Demand Planner/Forecast Parts see Forecast Part Info.
Inheritance End: The end date to which historical data is to be inherited. This field can only be defined in IFS/Demand Planner/Forecast Parts see Forecast Part Info.
Market Segment: The ID of the part's market segment.
Phase In: The phase in date for the part. The user is not allowed to set any forecast quantity for the part, before the period which contains this date. E.g. If using monthly period version and phase in date is 01.12.2003, the user is not allowed to enter a forecast for the part in the periods before December 2003 (2003-12). See Authorize Intervals.
Phase Out: The phase out date for the part. The user is not allowed to set any forecast quantity for the part, after the period which contains this date. E.g. If using monthly period version and phase out date is 01.12.2004, the the user is not allowed to enter a forecast for the part in the periods after December 2004 (2004-12). See Authorize Intervals.
Parts in group: Displays the number of items in the group when working on any group level.
Forecast From: The first period in the current forecast.
No of Forecast Periods: The number of periods in the current forecast.
Sum Forecast: The sum of all the period values of Adjusted Forecast in the current forecast. The unit depends on the choice made in Select Unit.
Forecast Changed By: The user who last changed this forecast part. This field is reset when a new forecast is created.
Inventory Unit: The current part's inventory unit.
Inventory value: The current part's selected inventory value. The possible values are:
Net Weight: The current part's weight. Read from inventory part.
Part flow value: This is a 'free' value field on the forecast part. This allows to individual values for a part on different flows.
Planned Sales Price: The current part's planned sales price, this price is normally fetched from IFS/Costing. The system can also use other user defined functions to read this value, so the value does not necessarily need to come from costing. See Registry setup for details of how to decide on this value and how to retrieve it.
Expected Demand Size: The expected demand value for the next period if demand occurs, and the field Inter Arrival Time is the average period limit between the demand periods. This field is equal to the forecast for all models except Croston's.
Inter Arrival Time: This field is the expected average period time between demand occurrences.
MAE: Mean absolute error, a widely used measure of forecast error magnitude. A detailed description of forecast measurements can be found in Forecast Measurements. Note. When looking on group level, this measure will be expressed in the same unit as what has been selected from the Select Unit combo. The lines used for the calculations are the ones shown in the forecast graph. If you are looking at parts, then this number is always expressed in inventory units and the calculations are based on the real demand statistics (demand in inventory units).
Value of MAE: Mean absolute error in cost price. A detailed description of forecast measurements can be found in Forecast Measurements.
MAPE: Mean absolute percentage error. A detailed description of forecast measurements can be found in Forecast Measurements. Note. When looking on group level, the calculations are done on the lines shown in the forecast graph. If you are looking at the forecast in another unit (Select Unit) than inventory units. The calculation of MAPE is done on shown demand/forecast in the graph. If you are looking at parts (part level) in other units, then this number is always calculated based on the real demand statistics (demand in inventory units).
WMAPE: Weighted mean absolute percentage error. A detailed description of forecast measurements can be found in Forecast Measurements. Note. When looking on group level, the calculations are done on the lines shown in the forecast graph. If you are looking at the forecast in another unit (Select Unit) than inventory units. If you are looking at parts (part level) in other units, then this number is always calculated based on the real demand statistics (demand in inventory units).
MSE: Mean squared error. A detailed description of forecast measurements can be found in Forecast Measurements. Note. When looking on group level, the calculations are done on the lines shown in the forecast graph. If you are looking at the forecast in another unit (Select Unit) than inventory units. If you are looking at parts (part level) in other units, then this number is always calculated based on the real demand statistics (demand in inventory units).
ME: Mean error. A detailed description of forecast measurements can be found in Forecast Measurements. Note. When looking on group level, this measure will be expressed in the same unit as the unit selected from the Select Unit combo. And the lines used for the calculations are the ones shown in the forecast graph. If you are looking at parts, then this number is always expressed in inventory units. And the calculations are based on the real demand statistics (demand in inventory units).
PVE: Percentage variation explained. A detailed description of forecast measurements can be found in Forecast Measurements. Note. When looking on group level, the calculations are done on the lines shown in the forecast graph. If you are looking at the forecast in another unit (Select Unit) than inventory units. The calculation of PVE is done on shown demand/forecast in the graph. If you are looking at parts (part level) then this number is always based on demand/forecast in inventory units.
Tracking Signal: This error measure is the sum of errors divided by the sum of the absolute errors. Instead of using the sum, an exponential smoothing formula is used, employing Delta as the smoothing constant. A detailed description of forecast measurements can be found in Forecast Measurements. Note. When looking on group level, the calculations are done on the lines shown in the forecast graph. If you are looking at the forecast in another unit (Select Unit) than inventory units. The calculation of Tracking Signal is done on shown demand/forecast in the graph. If you are looking at parts (part level) then this number is always based on demand/forecast in inventory units.
Adjustment Factor: Tells how much better the your forecast (historical forecast) are compared to the mathematical forecast (explanation of forecast), 0=equally good, -10 your adjustment 10% poorer than the mathematics 10 you adjustment 10% better than the mathematics. A detailed description of forecast measurements can be found in Forecast Measurements. Note. When looking on group level, the calculations are done on the lines shown in the forecast graph. If you are looking at the forecast in another unit (Select Unit) than inventory units. If you are looking at parts (part level) then this number is always based on demand/forecast in inventory units.
Theil's: Theil's U-statistic provides one approach for evaluating the performance of a forecast technique. This statistic compares the accuracy of the forecast technique prediction with a baseline in which the forecast is simply set to the historical demand from the preceding period. A Theil's U-statistic of 1.0 means that the forecast technique is performing as well as the baseline forecast. A value of less than 1.0 means the technique is performing better, while a value less than 1.0 means its performance is lower than that of the baseline forecast. A detailed description of forecast measurements can be found in Forecast Measurements.
R Squared: This is the coefficient of determination, and this is the proportion of variance accounted for (explained) by the regression (when you are using the regression forecast model ) or by the explanatory variables when using the multiple regression model. When you are using any other forecast model this field is 0.0. This is a number between 0 and 1.0, this number describes the proportion of the variation explained by the regression. A detailed description of the calculations of B Bar Squared can be found in Forecast Measurements.
R Bar Squared: This is also often called adjusted R Squared, or adjusted R2. This is R Squared adjusted for the degrees of freedom. R Squared will always increase when you add another explanation variable, this figure will increase only if the regression variable added give an increased explanation to the forecasted variable. If the added variable does not give an added value then the R Bar Squared number will be reduced. This number can also be negative. This number can be useful when doing stepwise regression (you are trying to build a regression out of many different explanation variables and are trying to find the most appropriate ones for the current forecasted part). When you are using any other forecast model this field is 0.0. A detailed description of the calculations of B Bar Squared can be found in Forecast Measurements.
Season Correlation This is the correlation between the parts seasonal profile and the part's 'real' seasonal pattern, when looking at the historical demand. A positive number indicates a positive correlation between the two seasons. This means that this parts season profile is a good match the higher the number the better the match. If the number is negative then the parts seasonal profile and its seasonal selling pattern is a mismatch. Correlation is a number between -1 and 1.
Forecast Comment: A free text field. This field can be moved around and resized like the select tool bars.