five-samples-per-dimension-ml-rule-of-thumb

IN premise — summaries/2026/08/24/wiki-Curse_of_dimensionality-chunk-2.md

Created 2026-08-24T17:11:08+00:00

A frequently cited practical rule of thumb in machine learning requires at least 5 training examples per dimension for reliable generalization.

Summary

If a dataset has many features, you need roughly five training examples for each one before a model can actually learn patterns rather than just memorize the data. This sets a practical floor on data volume: below that threshold, the system should treat any model output as unreliable and flag generalization risk accordingly.