An open library for your next question. Public pilot
Executable Science
Log inCreate account
← All resources
Datasets

Connectionist Bench (Sonar, Mines vs. Rocks)

Acoustic-return features for investigating uncertainty in a small binary detection benchmark.

Visit the original source ↗

External resource. No execution or independent verification is claimed here.

A QUESTION TO TAKE FURTHER

How variable is sonar classification across fixed stratified splits with train-only preprocessing?

After qualified ingestion and an immutable evaluation plan, recompute confusion matrices, classwise metrics and uncertainty from held-out predictions and frozen labels; fit preprocessing inside training partitions. Successful reproduction would not validate original measurements or causal interpretations.

What you could produce

  • Exact input manifest with data-file digests and a schema/rights audit.
  • Frozen evaluation protocol with split or group identifiers and an explicit baseline.
  • Bounded analysis report with recomputable metrics, uncertainty and unresolved findings.

Before you use it

  • Resolve exact data files and digests through qualified isolated ingestion.
  • Audit data schema, provenance and license conditions before designing a bounded analysis.
  • Freeze preprocessing, splits, metrics, runtime requirements and budget before execution.

Limits to keep in view

  • No dataset contents, dependency compatibility, runtime cost or scientific result was checked.
  • Grouping, timing or identifiers needed by the proposed protocol may be absent; report that limitation rather than inventing split labels.
  • Large or specialized sources may require a predetermined subset or a broader qualified runtime; no resource fit is established.
  • The qualified self-contained Python 3.13 pilot is not evidence that this dataset or its usual analysis packages can run.

Source and permission context

Attribute the supplied creators and UCI dataset DOI/source. Preserve license links and identify changes before any permitted data reuse.

Catalog listing reviewed. This review covers the description and source links displayed here.

Reviewed the original description of Connectionist Bench (Sonar, Mines vs. Rocks): Acoustic-return features for investigating uncertainty in a small binary detection benchmark. Its public UCI identity, supplied creators, scoped license observation and per-record page evidence match the stored official metadata. Listing approval concerns that description and source links; the recorded dataset license is not treated as a license for every linked artifact or as metadata rights clearance.

Reviewed 2026-09-14. Copying or adapting source files remains subject to their own terms.

data · CC-BY-4.0

This dataset's official page assigns CC BY 4.0 to the dataset. The observation concerns the dataset license statement; separate article, software and third-party rights were not inferred.

Inspect the license evidence ↗

Before copying source material

  • Data files and embedded or third-party notices were not downloaded or inspected.
  • The API's last-updated date is metadata, not an immutable dataset version or content digest.