Diagnosing Low Factor Coverage Caused by Missing Values on a Trading Date

Summary

This support discussion explains a factor-analysis failure caused by insufficient coverage of a derived price feature. The feature compares the close with the high-low range and then takes a one-period change. Although feature extraction returned data, the analysis module reported that only about 42 percent of raw factor observations were available, below its 50 percent minimum, because too many values were missing.

The response traced the problem to a date on which the factor was missing for all observations, possibly because of a data issue, and suggested excluding that date from the analysis. This is a troubleshooting example, not a study of the factor's predictive value. It does not establish why the entire date lacked values or whether excluding it is appropriate for other datasets; the underlying data and preprocessing should be checked before interpreting factor results.

Key ideas

• The factor-analysis tool rejects inputs when raw-value coverage falls below its stated threshold.
• A derived price-range feature had missing values for an entire date.
• The response suggests excluding that date as a possible workaround.
• The cause is attributed tentatively to data issues and is not independently established.

Tags

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