Conformal Prediction After Data-Dependent Model Selection: Methods and Applications

A 2.2 MB research dataset from figshare, last updated May 2026, authored by Ruiting Liang. It supports a paper on constructing valid conformal prediction sets after selecting a model to minimize set width, without requiring additional sample-splitting. The dataset includes synthetic and real-world examples to demonstrate the proposed methods.

Use Cases

• Benchmarking novel conformal prediction methods based on the described data-dependent model selection framework.
• Testing finite-sample validity guarantees for prediction sets based on the described synthetic datasets.
• Comparing prediction set width and coverage performance across different model classes as described in the paper.
• Reproducing the real data example application mentioned in the research description.

Strengths

• Includes a real data example alongside synthetic datasets for validation.
• Methods are described as efficient and admit finite-sample validity guarantees.
• Released under a permissive CC-BY-4.0 license for reuse.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The primary files are PDF and ZIP, suggesting the data may be packaged within documentation.

Provenance

Source

Ruiting Liang via figshare.

Collection Method

Likely generated for research validation, including synthetic data and a real-world application.

Time Range

Temporal coverage is not specified in the metadata.

Freshness

Last updated 2026-05-01 15:03:51; freshness should be verified.

Geography

Spatial coverage is not specified in the metadata.