Skip to content
All library documents

StockRanker Feature Importance and NDCG Evaluation in BigQuant

Article BigQuant
Read the original

Summary

This platform discussion addresses a reporting difference between older and newer versions of a StockRanker workflow. A user says the older version displayed feature importance and NDCG after a backtest, while those results were not visible in the newer version. A reply says feature importance is still available in the newer version and that NDCG requires labeled data for the prediction set. The exchange points to a separate guide for viewing feature importance, but does not reproduce its instructions. It offers no model results, explanation of the metric, or details about the ranking strategy, so its practical value is limited to troubleshooting result visibility and the labeled-data requirement for NDCG.

Key ideas

  • The discussion concerns missing feature-importance and NDCG outputs in a newer StockRanker interface.
  • A reply states that feature importance remains available in the newer version.
  • NDCG calculation requires labeled data for the prediction set.
  • The exchange directs readers to another guide for instructions on viewing feature importance.

Tags

Machine learningEquitiesStatisticsBacktesting

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