Diagnosing Binary Probability Outputs from XGBoost Classifier

Summary

The document reports a platform-specific issue in which an XGBoost classifier’s probability prediction function returns only endpoint values of zero and one. The user expected continuous class probabilities that could be used to rank securities, based on behavior seen in a local installation, but the platform output shown consists of the same hard, extreme probability pair across observations.

No solution, diagnosis, or follow-up evidence is included. The example raises a practical concern for quantitative modeling: probability outputs that collapse to binary extremes are not useful for probability-based ranking and may indicate a difference in model configuration, platform implementation, or data and prediction setup. The document does not provide enough information to determine which cause applies or whether the reported behavior is reproducible.

Key ideas

• The reported classifier output contains only zero and one probability values.
• The user wants continuous probabilities to rank securities.
• The document gives an example output but no explanation or fix.
• Platform configuration, implementation, or input differences remain possible causes.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.