Manifold analysis and the stability of cross-validation
The bibliography cites a manifold-based robustness analysis and a theory of cross-validation through stability. Both concern the geometry of the input space and the reliability of the estimate.
The requirement in plain English
The bibliography cites a manifold-based robustness analysis and a theory of cross-validation through stability. Both concern the geometry of the input space and the reliability of the estimate.
What the public record says
The bibliography entries name the works. The public view gives no summaries and no clause references.
Why it matters to you
A statistical estimate assumes that the samples behave like the population. Cross-validation stability is one way to test that assumption, and manifold analysis is one way to describe the region the data actually occupy. Together they answer a question a reviewer will ask: is the result stable enough to act on? A single split that produces a good number is not evidence of stability; repeated splits that agree are.
Evidence a reviewer asks for
- The cross-validation scheme, the number of folds and the repeats.
- The variance of the estimate across splits.
- The manifold or region description used to justify the sampling frame.
- A note on any split that produced a materially different result.
Checks to run this month
- Is the result stable across splits?
- Is the data region described, not assumed?
- Would a different split change the decision?
Dates
Papers published 2024 and 2025. FDIS registered 2026-08-21.
Cross-references
- A statistical approach to robustness
- Uncertainty, complexity and model confidence
- Write the sampling and perturbation plan
What this page does not state
The public view gives no summaries and no statement of how the draft uses the papers.