How To Fix ACE_ML017 - The class distribution ratio must be equal to or greater than 10%.

SAP Error Message - Details

• Message type: E = Error

• Message class: ACE_ML -

• Message number: 017

• Message text: The class distribution ratio must be equal to or greater than 10%.

What causes this issue?

Class Distribution Ratio = Records that contain revised accrual amounts
/ All records for data training.

System Response

The system issues an error message and will not allow you to continue with this transaction until the error is resolved.

How to fix this error?

Procedure for System Administrators

Error message extract from SAP system. Copyright SAP SE.

Smart SAP Assistant

• What is the cause and solution for SAP error message ACE_ML017 - The class distribution ratio must be equal to or greater than 10%. ?

  The SAP error message ACE_ML017 indicates that the class distribution ratio in your dataset is not meeting the minimum requirement of 10%. This error typically arises in the context of machine learning models, particularly when you are trying to train a model that requires a balanced dataset across different classes.
  Cause:
  The error occurs when one or more classes in your dataset have a representation that is less than 10% of the total instances. For example, if you have a dataset with 1000 instances and one class has only 5 instances, that class represents only 0.5% of the total, which is below the required threshold.
  Solution:
  To resolve this error, you can consider the following approaches:
  
  
  Data Collection: Increase the number of instances for the underrepresented classes. This may involve collecting more data or augmenting existing data.
  
  
  Resampling Techniques:
  
  Oversampling: Increase the number of instances in the minority class by duplicating existing instances or generating synthetic instances (e.g., using SMOTE - Synthetic Minority Over-sampling Technique).
  Undersampling: Reduce the number of instances in the majority class to balance the dataset. Be cautious with this approach as it may lead to loss of valuable information.
  
  
  
  Class Weighting: If you cannot balance the dataset, consider using class weights in your model training. This approach allows the model to give more importance to the minority classes during training.
  
  
  Change the Model: Some algorithms are more robust to class imbalance. Consider using models that can handle imbalanced datasets better.
  
  
  Data Transformation: If applicable, transform your data to create new features that may help in better distinguishing between classes.
  
  
  Related Information:
  
  Model Evaluation: When dealing with imbalanced datasets, traditional accuracy metrics may not be sufficient. Consider using metrics like precision, recall, F1-score, or the area under the ROC curve (AUC-ROC) to evaluate model performance.
  Documentation: Refer to SAP's official documentation for specific guidelines on handling class imbalance in the context of the machine learning tools you are using.
  Community Forums: Engage with SAP community forums or support channels for additional insights and shared experiences from other users who may have faced similar issues.
  
  By addressing the class distribution ratio, you should be able to resolve the ACE_ML017 error and proceed with your machine learning tasks in SAP.