|
Forecast Error Analysis dialog box |
|
Use this dialog box to report normalized mean square errors (NMSE) and confidence intervals for forecast errors at specified prediction steps.
Specifies the solution method to test. Select one of the solutions from list.
 | This box will only list solutions that are ready for use. |
Refer to the Solution Methods dialog box for additional information on this subject.
Output
Specifies the worksheet location for placing the results of this operation.
Select this option to place the results of this operation in a new document workbook.
Select this option to place the results of this operation in a new worksheet.
Select this option to place the results of this operation starting at the worksheet location specified in the box to the right of this option.
Refer to Specifying Output Locations for additional information on this subject.
Check this box to plot the results of this operation.
Input and Test Data
Displays the worksheet extent of data that can be used for the test as a reference. Use this number as a guide for specifying the test Start Position.
Displays the amount of data left for input after the test Start Position value is specified. Use this number as a guide for specifying the Test Size.
Indicates the index in the data where the test is to begin. Specify any index smaller than the number indicated by Total Available.
| The index indicated by the Start Position value is relative to the first output position in the data. In this context, Start Position zero corresponds to the first output vector in the data. |
 | It is a good practice to specify a Start Position value that includes a substantial number of points when testing local methods. |
Indicates the length of the test. Specify any length less than or equal to the number indicated by Available For Use.
Refer to Data and System Models in Predictive Systems Lab for additional information on this subject.
Test Context
Indicates the minimum number of time steps into the future to predict for from the Start Position in order to gather the required error information. This value must be equal or greater than 1 and equal or smaller than Max. Steps.
Indicates the maximum number of time steps into the future to predict for from the Start Position in order to gather the required error information. This value must be equal or greater than Min. Steps and equal or smaller than the total length of the data minus the solution method overhead requirement.
| This value may impose restrictions to the Start Position and the Test Size. |
Indicates the confidence level for which you want the command to calculate confidence intervals. Confidence intervals are calculated by approximating the inverse cumulative distribution function of the residuals by assuming, first, that the number of samples constitute 98% of a larger unknown population, and by approximating the resulting cumulative distribution with the logistic function under least squares methods. Type in this box a value greater than zero and smaller than one.
| Please note than non-iterative models will have equal results for all prediction steps for global methods. For local methods, indicating prediction steps greater than 1 for these models will have the effect of freezing the nearest neighbor database to the number of prediction steps less one in the past of the current prediction point, so the results will differ as the pool of nearest neighbors to compute the predictions vary. |
Refer to Data and System Models in Predictive Systems Lab for additional information on system models.
Closes the dialog box and carries out this operation.
Closes the dialog box without carrying out this operation.
Sample Output
Sponsored
Try Predictive Systems Lab
Need Help?
Contact support