Fixing Rolling Training Errors in a Convertible Bond AI Strategy

Summary

This troubleshooting note addresses errors that arise after adding a rolling-training module to an AI strategy for convertible bonds. It points to a revised notebook and identifies two implementation changes associated with the fix. First, the rolling backtest should not use the data source directly as the inputs for the intermediate training and testing instrument fields. Second, the workflow uses an updated missing-value handling module.

The note is a concise implementation hint rather than a full diagnostic guide. It does not include the error message, code, explanation of the required intermediate data transformations, or evidence comparing results before and after the changes. Readers can take away that rolling training may require preparing instrument inputs separately and aligning missing-value handling with the current module, but they will need the referenced notebook or their own debugging to determine the exact modifications for a particular workflow.

Key ideas

• The reported issue appears after adding rolling training to a convertible bond AI strategy.
• The rolling backtest should not use the raw data source directly for intermediate training and testing instrument inputs.
• The suggested workflow uses a newer missing-value processing module.
• The note omits the error details and exact transformations, so it is not a complete debugging procedure.

Tags

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