Documentation Application Tour Get Involved

Statistical Limitations

The deterministic 60-example fixture is designed to exercise software paths, not estimate field performance. Its balanced labels, synthetic presentation, small size, and repeated construction limit external validity.

Phase 9 records sample and class counts, but does not invent confidence intervals or significance tests where assumptions are unmet. Research reports explicitly disclose missing statistics, leakage warnings, and data origin.

 _   _           _     ____
| | | | ___ _ __| |__ |  _ \  _____   __
| |_| |/ _ \ '__| '_ \| | | |/ _ \ \ / /
|  _  |  __/ |  | |_) | |_| |  __/\ V /
|_| |_|\___|_|  |_.__/|____/ \___| \_/