Troubleshooting Advanced Predictions Issues

Use the information in this section to resolve issues related to Advanced Predictions.

Applies to

FreeForm, Planning (all application types), Sales Planning, Strategic Workforce Planning.

Frequently Asked Questions

Review these frequently asked questions to see if they resolve your issue:

  • What is the minimum amount of historical data required to run Advanced Predictions?

    Advanced Predictions require a minimum of twice the length of your prediction range as historical data. For example, if you need to predict 15 months into the future, at least 30 months of historical data is required.

  • What drivers should I use for Advanced Predictions?

    Drivers that are strongly correlated with the target account would produce good prediction results. Both internal and external (macro-economic data) drivers can be used for good prediction result.

  • Do I need future data for driver members when running Advanced Predictions?

    Yes, you need future data for drivers. In case you do not have future data for drivers, select Predict missing input driver values when configuring the Advanced Prediction job to predict values where input driver values are missing. Missing values are predicted using statistical forecasting (univariate predictions), and are used for ML model training.

  • Can I run Advanced Predictions for a Dynamic Calc parent account member?

    Yes. Advanced Predictions for Dynamic Parent accounts is available starting with the 25.11 update.

  • What user roles are required to create and run Advanced Prediction jobs?

    To create and execute Advanced Prediction jobs, users need the Service Administrator or IPM MANAGE role in Planning.

  • Is it possible to schedule Advanced Prediction jobs to run automatically?

    Yes. Advanced Prediction jobs can be scheduled using the Data Integration pipeline in Planning or FreeForm.

  • How can I evaluate the accuracy of Advanced Predictions?

    Planning provides built-in accuracy metrics including MAPE (Mean Absolute Percentage Error) and RMSE (Root Mean Square Error). These are available as a downloadable report on the Prediction job results panel for each job.

  • Can planners override or adjust Advanced Prediction values?

    Oracle does not recommend overriding prediction data as it would cause traceability issues. Prediction data can be used as a baseline and if any adjustment is needed, it can be done though an adjustment member in Data Element or any representative dimension that provides data lineage.

  • Do Advanced Predictions handle outliers or anomalies in historical data?

    Yes. The prediction engine includes automatic outlier detection and smoothing. It identifies unusual spikes or dips that deviate significantly from historical norms and smooths the data based on the selected configuration for outlier adjustment.

  • What happens if there is missing or sparse data in the historical time series?

    Advanced Predictions handle sparse data through imputation techniques. Where gaps exist, the engine uses surrounding data points to estimate missing values before training the model. However, users can specify different options for missing data treatment.

  • Do Advanced Predictions work with BSO and ASO cubes?

    Advanced Predictions work with both BSO and ASO cubes.