What is the maximum number of levels for a class variable that triggers it to be marked REJECTED?

In SAS Enterprise Miner, a class variable is marked as REJECTED when it exceeds a certain number of levels, which can diminish the effectiveness of model training or increase complexity unnecessarily. The threshold for triggering this REJECTED status is set at 20 levels for a class variable.

When a class variable has more than 20 levels, it may lead to issues such as overfitting, increased computational time, and difficulty in interpreting the model. Fewer levels generally provide clearer insights and make the model more manageable. Thus, if a class variable contains more than 20 distinct levels, it is flagged as REJECTED to prevent potential issues during analysis and modeling.

This policy ensures that the modeling process remains efficient and interpretable, which is crucial for effective data science practices. The maximum set at 20 levels is thus a practical guideline used in SAS Enterprise Miner to maintain the integrity of data analysis tasks.