Defining forecasting parameters


This topic describes how to define forecasting parameters for a time forecasting model scenario, which is part of Adding-a-time-forecasting-model-scenario.

To define forecasting parameters for a time forecasting scenario

Warning

Important

To define forecasting parameters marked with 🔵️, click Advanced.

On the Add a new scenario page, specify the following:

  1. Name: Type a name for the scenario.
  2. Description: Type a short description of the scenario.
  3. Time resolution: Select the series interval (hour, day, or month).
  4. Analysis period type: Select the amount of time for which data is analyzed when formulating a prediction.
    • Manually specify: If you select this option, for Period of, type and select the number of hours, days, or months prior to the scenario execution date for which data is analyzed
    • Use time filter: If you select this option, for Time filter, select an existing time filter. You can further modify the selected time filter by clicking Edit.
  5. Forward forecasting period: Select how far into the future the series is forecast, with respect to the end of the analysis period:
    • Specific period: If you select this option, for Forecasting period of, type a number and select day or month.
    • Until date: If you select this option, for Until date, select a date from the calendar.
    • Until end: If you select this option, for Until the end of current, select month or year.
  6. Algorithm: Select an algorithm that determines how series are projected over time. For a list and descriptions of supported algorithms, see Supported algorithms.
  7. Include committed reservations: Select True or False.
  8. Threshold type: This field indicates the threshold type.
  9. 🔵️ Apply threshold only in the future: Select True or False.
  10. 🔵️ Threshold direction: Select one of the following:
    • Automatic
    • Saturation achieved when forecast is greater than threshold (Best Lower)
    • Saturation achieved when forecast is lesser than threshold (Best Higher)
  11. If you selected Manual thresholds when you created the time forecasting model, then, for Thresholds value, select one of the following:
    • Same value for all series: If you select this option, then, for Value, type a percentage value for upper thresholds of all series.
    • Specific threshold: If you select this option, a Threshold box is displayed for each series you selected when you created the time forecasting model. For each threshold, type a percentage value for the upper threshold value of that series.
  12. If you selected Manual thresholds given by other metric when you created the time forecasting model, for Threshold scale factor, type the percentage to use for scaling the threshold.
  13. Show reliability indicator: Select Yes or No.
  14. Threshold scale factor: Enter a value for the threshold scale factor.
  15. 🔵️ Update policy: Select when results are updated:
    • Manual
    • Automatic on page access
    • Automatic on time filter, system and business driver changes
  16. Click Save or continue with Defining-algorithm-options.

Supported algorithms

For more information about algorithms, see KB102 - Time Forecasting models.

Where to go from here

After you have defined forecasting parameters for a time forecasting model scenario, you can continue with Defining-algorithm-options.

 

Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*

BMC Helix Continuous Optimization 25.1